East Coast Sound Modulator Circuit?

[BrianBurnsSWL]

Hello All,

I've got a long term project going of a rig with an 813 in the final that I would like to plate modulate with Hi-Fi audio. Can anyone suggest a source for a tube modulator circuit, capable of 100 watts or so, that is designed from the beginning with high quality audio in mind?

Thanks a bunch,

73

Brian Burns.....soon to be K6UCD once again

[Opcom]

Any of the modulators used in 250 Watt broadcast transmitters would be worth looking at. The tubes and high voltages/currents can easily be downsized for "QRP" applications.

For example, the RCA BTA-250 modulator uses 828 tetrodes driven by 6J7's in AB1, but choose any 250W transmitter that pleases you. In particular pay attention to the way audio feedback is done. It's different than most hi-fi amps.

AF input: +16dBm
AF Distortion: less than 3% RMS from 50-7000 cycles
Hum and noise: 60dB below 100% modulation

Modulator plate voltage: 1650
Modulator plate current (100% mod) 260mA
RF stage input 1500V @ 260mA  (390W)

If those specs are hi-fi enough. Other/later transmitters might do better.

Otherwise try something like the Altec 1570B audio amp. 1% distortion using 811A's and the only real frequency restriction is the output transformer and maybe the driver choke at extremes.

I tried to suggest things that would allow a fairly common modulation transformer to be used. - and if using a ham or communications one and wanting better low frequency response, see the posts on transformer power vs frequency, and/or oversize it 2-3x !

[WD5JKO]

Can anyone suggest a source for a tube modulator circuit, capable of 100 watts or so, that is designed from the beginning with high quality audio in mind?

   Brian,

   The term "High-Quality" on amateur AM, and in the living room audio systems are quite different. For amateur radio, running AM we have to consider the receiver at the other end might be using a 6-8 Khz bandwidth, so that means audio frequencies above 3-4 Khz are just not heard, and if present the extra transmitted bandwidth just widens your signal which interferes with other stations adjacent to your carrier.

   I got a good lesson from this about 25 years ago from Don, K4KYV. I had a Globe King 500B, and I built a solid state audio driver for the 811 modulators. Man I worked on this thing really hard. I got that king to fully modulate from 50 hz to 10 Khz, and had flat frequency response, low distortion, and the whole ball of wax since in my mind this was the best modulator ever built. Then my first QSO was with Don on 160M, I had a good signal, but he had trouble copying me as my audio was all muddy from being un-equalized. I also had a condenser Mic that was high quality with a broad flat response.

 Lesson 1: We need some sort of equalization. Sure include the bass, but consider that receiver at the other end. Do you want to sound like a "woofer" with no mids or highs?  Look at this old article from W2WLR:
http://www.amwindow.org/tech/pdf/eam.pdf

  Lesson 2: We need to achieve a high average modulation percentage. This means we need to get our audio polarity optimized such that we can peak > 100% upward, maybe 125% as a goal without clipping 100% at the baseline. To pull this off we need extra audio power, and a modulation transformer turns ratio appropriate to the task.

  Lesson 3: If you plan to run a lot of global negative feedback (NFB) to lower your distortion, then make sure you never overload (clip) the modulator. With lots of NFB (say 20db), when you clip the loop opens at the clip point, and the gain increases 20 DB making the overload much more pronounced. If the modulator is class Ab1 with R-C coupling, and big fat capacitors to pass bass frequencies, then during that overload there will be momentary grid current on the modulator tube(s), and the grid current will result in a DC shift in the operating point of the tube overloaded. This shift has to decay at an R-C rate, and with a big R (plate load R of the driver), and a big C (coupling capacitor), this could take a half second. So a one syllable overload can make a huge nasty noise out of your modulator. So less NFB, faster R-C time constant, or transformer coupled interstage transformers do have their advantage. For this reason, I personally limit global NFB to 3-6 DB, and make another loop around the audio driver if I'm driving class B grids.

Many will disagree with me, and so be it. What we do in the living room stereo needs to be different from the AM modulator.

  So consider whether your modulator has tubes, or solid state. with tubes, do you want triodes, or screen grid tubes? For 100 watts audio, I'd shoot for 150 for headroom. Here a pair of 811's will do nicely, as would 4 - 6 type 807's in class Ab1. Also look at the new kid on the block, the KT-120:

http://www.audiobanter.com/showthread.php?t=135926

These are complex tube living room amps to ponder, but remember the W2WLR lesson:
http://www.turneraudio.com.au/poweramps.html

Jim
WD5JKO



 

[flintstone mop]

Hi Brian
I hope we can work each other soon when you get the TX on the air.
Jim brings up good points. It's going to depend on how anal you want to get with the modulator.
My favorite TX for plenty of pleasant audio was the RA 250. Four 2A3's driving the two 810 modulator tubes. Triodes much easier to deal with. I do not like 4-400's.
If you like to take little short cuts, try this: You have built your modulator with it's final tubes. Bias supply built and plate supply. You have a reasonable match with the output tubes and your RF final (mod iron and Heising hook-up). Drive this modulator deck with a quality old fashioned tube P.A. amplifier. Get a working audio output transformer and hook it up backwards. The plate leads connected to the grids and the 8 ohm side connected to this P.A. amp 20 watt amplifier (usually enough audio). The center conductor lead can be the wire that brings needed bias to the output tubes in the modulator.
A small mixer is needed to properly terminate a good quality microphone. Some mixers include EQ and even some mic processing. The idea here is the microphone has to be properly terminated. A $99 Marshall condenser mic. is beautiful!!
Looks like a lot of stuff, but others will have different approaches. There is no short cuts. You cannot plug a good quality mic directly into a DX100 type of circuitry. If you want real East Coast sound that is nice balanced audio, then it gets a little complicated.
The heavy bass folks is not reality and usually hard to copy when band conditions get crappy. And some folks have receivers that are 4kc bandwidth and that will kill any intelligibility out of the received signal. Paul, VJB has a good solution with an EQ setting of a smiley face. Boosting gently starting around 3kc and a sharp cutoff around 8kc.
If you want to get simpler, then use the mic mixer, EQ, mic processing, small amp and design your RF amplifier to be screen modulated. Frank, The Slab, has a really nice TX that has beautiful TX audio using screen modulation of his finals.
That's it from here.
Lotta info from this Forum.
Fred

[K5UJ]

One of the biggest myths in AM (and "ESSB") audio is the idea that dramatically boosted low frequencies, especially those below 100 hz are vital to the achievement of "broadcast quality" audio.  Nothing could be further from the truth.  Listen to any AM bc station and you'll hear people who sound no different from what they sound like in person.  There may be some alteration due to compression, but there's never a frantic attempt to get someone who sounds like Sean Hannity to sound like James Earl Jones.   But some hams are obsessed with the idea that some processing boxes will give them what nature has not, a deep rich baratone voice.   I've heard stations that are basically unintelligible because most of their audio power is concentrated in these power wasting very low frequencies.  A few with naturally low voices should actually attenuate those frequencies some.   Most male voices are naturally in the mid-range and highs and are lucky for that reason and should exploit that with a flat response curve from around 125 hz up to 3 or 4 kc before rolling off.  This will produce a natural sound that is easy to copy--the real definition of "broadcast quality."

[steve_qix]

Hello All,

I've got a long term project going of a rig with an 813 in the final that I would like to plate modulate with Hi-Fi audio. Can anyone suggest a source for a tube modulator circuit, capable of 100 watts or so, that is designed from the beginning with high quality audio in mind?

Thanks a bunch,

73

Brian Burns.....soon to be K6UCD once again

Hi Brian - Long term project, eh  

Anyway, we need some more info.  On your 813 - how are you going to run it?  An 813 will do 2000V at 200mA - 400 watts input - no problem.

2nd - how "East Coast Sound"-ish do you want to be (in other words, how good do you want your modulator to be)?

Excellent: Pulse width modulation (high power, high level modulation, no modulation transformer)  (or screen grid modulation - low power but capable of superb quality)

Very Good: 1:1 ratio modulation transformer with something like 572Bs as modulators; direct coupled audio driver; negative feed back; high level negative peak limiter; near-200% positive modulation capability.  Circuits are available.

Pretty Good: Some other ratio modulation transformer (1.5:1, 1.7:1, etc.) with 572Bs; direct coupled audio driver; negative feed back; high level negative peak limiter; positive peaks limited by transformer ratio.  Circuits available.

Good: transformer coupled audio driver of an otherwise good design (maybe an external solid state audio amp and a good quality transformer); high level negative peak limiter; limited or no negative feed back (due to phase shift problems in the driver transformer)

Not so good: so-so designed transformer coupled audio driver

Bad: under powered modulator


Anyway, what are you thinking about for power, and do you have modulation iron in hand?

[BrianBurnsSWL]

Hello All,

Just a note to say thanks very much for all the excellent replies! Give me some time to go through them all carefully. I'm a real neophyte at AM, having been exclusively a CW operator back in the day, so I'm trying to learn all I can.

73

Brian

See my website at :  www.lessonsinlutherie.com

Toll Free: 866-214-9472

[BrianBurnsSWL]

Hello Steve,

I thought that I should answer your questions.

Yes, "long term project" sigh...Having chosen the artsy-craftsy life (guitar maker), which requires taking the vows of poverty, progress on the rig will be largely dependent on the availability of money for parts.

The rig I have in mind uses one 813 in the final, running class C at probably 1500 V on the plate at around 150 ma. I want the tube to be just loafing along, and last a long time.

As for how good do I want the modulator to be, I would like it to be good enough that the other guys find it easy to copy, and pleasant to listen to. BTW I use audio spectrum analysis software to voice my instruments, so that is available for  checking out what is really happening.

No "modulation iron" in hand, though I found a sympathetic engineer at AnTek who is looking into what they have in their toroidal transformers that might work. They have a helluva deal for a power transformer--800-0-800 @ 900ma for $98, and it even has a filament winding!

I'm one of those Luddites that wants to build exclusively with tubes, so I suppose that leaves out pulse width modulation.

Again, thanks very much everyone, for all the thoughtful responses. I'll be following up on all the links, and am printing out copies of everything for reference.

73

Brian

See my website at :  www.lessonsinlutherie.com

Toll Free: 866-214-9472

[steve_qix]

My first pulse width modulator was all tubes 

[Opcom]

pulse mod with tubes, bet it was clear-sounding. DC to daylight. That makes my point.

A.) A modulator is not to be used as the limiting stages for frequency response.

B.) Nothing should be put into a modulator or audio power amplifier that it can not reproduce faithfully at full power.

There are a couple of ham stations around here that sound like they are talking through a carpet or a woofer and I bet the modulation scopes on them show perfect 100% modulation.

The bass part of their voices, for whatever reason (mike & load, bass boost, etc) is causing solid 100% modulation but the stuff above 300Hz is causing only 10% as a guess. The operator ought to EQ it.
For those who don't want a computer in the radio room, there are cheap hardware EQs that include a crude audio spectrum analyzer display of sorts.

I hate to say it but I don't like working one guy in particular because I can hardly understand what is being said. Operator does not believe anything is wrong. I dread when his turn is before mine. Sounds like umunuhamumbleomaamonuounamuuobblemm.

[flintstone mop]

Hey Brian
We do unusual things here. You CAN use a power transformer as a modulation transformer.
Here's a link on how to do it. I did it and was surprised. Your toroid transformer would be a good candidate.

http://www.amwindow.org/tech/htm/tvtomod.htm

It seems like you might want to look into screen modulation.......much less money involved for iron. PWM the parts count increases and more involved to set up.

[K5UJ]

The bass part of their voices, for whatever reason (mike & load, bass boost, etc) is causing solid 100% modulation but the stuff above 300Hz is causing only 10% as a guess. The operator ought to EQ it.

The only thing you can to with those numb skulls is run the rx audio through a simple equalizer, chop off everything below 200 hz and boost everything above it.  I have to do that occasionally with one or two stations I listen to.  It's amazing how they go from being unintelligible to being at least moderately understandable.   I'd also say you could record them and play it back but people like that are so into denial that they'd probably just tell you your rx is faulty, or you deliberately altered their signal or something.
I'll bet they have never monitored their tx audio with a flat response wide passband rx directly off the air.

[KM1H]

If you play it back without saying anything they will likely come back saying YOU have a problem since they cant understand it

After awhile even the thickest numbskull will get the hint.

[flintstone mop]

Some folks cannot determine what to tweek or cannot monitor themselves and take a good guess.
They might rely on another operator with a Narrow bandwidth receiver or lousy monitoring and cannot give accurate information to the operator trying to adjust.

Basically what we're doing here is almost like a Bcast studio. A microphone properly terminated, 8kc BW receiver with decent audio, and hi-fi monitoring , speakers or headphones.
If your radio sounds super listening to music on AM ( hard to find these days) or a Short wave station playing music, then you have a good monitoring system. That is the beginning.

[BrianBurnsSWL]

Hello Fred,

Speaking of monitoring, has anyone had any experience using a computer "'scope" program as a modulation monitor? There are a couple of free ones that I know of, "Visual Analyser" being a very good one.

Cheers,

Brian

[WBear2GCR]

Brian,

Visited your site, very impressive work!

Regarding your plans to build a rig. There are a host of ways to skin your cat. However, this: http://www.amwindow.org/tech/htm/813/813.htm  is a proven way. A variation of this rig has been built buy two hams I know personally that are local, and it worked out for them very nicely.

My suggestion to you is to not build a single 813, but to build a pair. And I would build in the higher limit voltage power supply and dial it back with a variac if you want to "loaf".

Building a pair of 813 is just as easy as a single, and you get a bunch more powah! Out there on the west coast, you're not a densely packed in ham-wise as we are here on the east coast, so you will want to be HEARD and for that there is no substitute for powah! 

Of course this is presuming a good antenna first. No ant, no signal.

If a rack size transmitter is an issue for you, then you don't want tubes, you want a Class E rig... but once ur building a full size (4' rack or 6' rack) transmitter, there's no point in not packing in as much juice as is practical.

Go for the 2 x 813 finals!

                      _-_-bear

PS. the modulation side is easy enough - and you can always pull the modulator chassis and use something else. For example you can start with the 2 x 813 RF deck modulated by an outboard audio amp via a cheap transformer, then build up another type of modulator and swap it in, IF you ever want to!

[BrianBurnsSWL]

Thanks for the kind words about my website! It looks like that bear has a guitar neck in his hand (?)

The other day I printed out the schematic for that four 813 rig because it has a Pi-net output.

Since I'm on a budget that is tighter than tight, thanks to my choice of an artsy-craftsy life, I need to keep an AM rig down to what I can reasonably afford.

I have a character flaw of taking on more than I can possibly accomplish, and at age 72 I'm finally trying to come to grips with the reality of the situation (;->)...

A wise friend once told me that every project will "take longer than that, and cost more than that".

Cheers,

Brian

[steve_qix]

Ok, if you really have no money for this sort of stuff (and believe me, I have BEEN THERE, and still in fact spend very little on ham radio), you can still do it.

1) MAKE EVERYTHING YOU CAN.  That means winding coils, using existing metal that you have or going to Home Depot and buying angle stock (which is very cheap) and pieces of aluminum and putting it together.

2) Take an inventory of all of your radio parts (within reason).

3) If you don't have expensive parts such as modulation transformers, filament transformers, etc.  switch to a design that doesn't use those things.  Consider class E. If you are not picky about size or what it looks like, you can build these things on the cheap.

4) Use Ebay, Craigslist, flea markets for sources of parts.  I have found some AMAZING deals using all these sources.

5) Recycle - Take old equipment apart to free up components for new projects.

6) Don't discount your dump or "recycling center".  Many towns have also "give and take" sheds - these can be goldmines for parts if you get the right stuff.

If you're really smart and really good at building, reusing, and scrounging parts, you can build a 400 watt transmitter for probably $150.00 - maybe LESS if you're REALLY good.  

Albeit, times are not quite what they were, but back when I first got on the air (at around 16 years old), I had absolutely no financial resources whatsoever.  

But, despite this I was able to build a solid 100 watt transmitter *AND GET A RECEIVER* for $0.  Everything came from the dump - an old Magnavox radio from the 40s that covered 160 and 75 meters as a receiver, a TV chassis stripped of all parts, a couple of horizontal sweep tubes in the RF amp and a bunch more in parallel as a series modulator (NO modulation transformer needed), TV power transformers in series for the power supply - broadcast variable caps for tuning and loading, hand wound coils - you get the idea.

The very first transmitter was a screen modulated horizontal sweep tube, but that was eventually replaced by the one described above.

Just pointing out that you can do this on the CHEAP CHEAP CHEAP.  You just have to be more resourceful than the guys with the money  

[BrianBurnsSWL]

Hi Steve,

Thanks for the thoughtful reply. I do have some money for ham gear, just not much. I've pretty well used up my current budget on an HQ-150 that I'm having re-furbished, that sort of work being "above my pay grade". The ham budget refreshes itself, at a trickle.

That series of Hammarlund receivers was what I lusted after back in the day, and after 57 short years I'm getting one! Talk about "wish fufillment"!

This transmitter that I'm contemplating building (QST, July 1951) has home brew plug-in coils, and I'm figuring on making a Pi-section output tank with a shorting bar for the final. Lenny W2BVH, has very kindly sent me a couple of used 813's, and I am very much the scrounger and DIY guy when it comes to making things.

Cheers,

Brian

 

[steve_qix]

Gee, if I think about it, one of the cheapest GOOD receivers one could get is ANYTHING that tunes short wave reasonably (like an old tube radio chassis from the 40s or 50s) and has a 455kHz IF.

Then build a soft rock (about $15.00), hook it up to the IF (you may have to build an interface amplifier or something).... get an old computer (they are throwing out Pentium 4s - perfectly adequate for running the FREE PowerSDR software - and you've got an awesome receiver for small bucks     Of course, my definition of low budget is VERY low (under $100 if possible)   

[KC2ZFA]

This transmitter that I'm contemplating building (QST, July 1951) has home brew plug-in coils, and I'm figuring on making a Pi-section output tank with a shorting bar for the final.

I suggest you look at the 813 xmtr in QST June 1954 (pg. 16 on).

Peter

[BrianBurnsSWL]

Hello Peter,

Thanks for the suggestion! I had looked briefly at the QST June 1954  article, but hadn't noted the Pi-section output tank. I certainly prefer Pi-section to the L/C circuit of the transmitter in the QST for July 1951.

I gave the 1954 article only a short look because I want to stay as simple as possible, and plug in coils are a lot easier than band switches.

BTW, there is a screen-grid modulator in the October 1951 QST that is the companion to the QST July 1951 transmitter. If screen grid modulation can produce high quality at the loss of some power, that may be the most economic way to go.

Cheers,

Brian

[steve_qix]

Screen modulation produces superb quality, but at MUCH lower power for the same tubes and power input.  The efficiency is 33% (or less, depending).  But, it is certainly CHEAPER to build it - no mod transformer, no audio driver.

A *PAIR* of 813s running screen modulation will put OUT about 110 watts or so running at full power INPUT, which, in this mode would be about 350 watts.  The transmitter will be simpler to build than most other designs.  Run the highest voltage possible - 2500V if you can get it - in this mode.

Does the transmitter *have* to cover above 40 meters?  If not, MAKE a coil out of #10 copper and use a clip lead to change bands (I'm trying to save you money here).  Use a PI section output - much simpler.  On a very limited budget, things like high voltage bandswitches are luxuries 

Don't worry about finding a circuit.  If you're really serious about building this thing, get the parts, do the metal work, mount the major components and I will commit to providing a suitable circuit.  That's the easy part.  The metal and mechanical work is the hard part.

[BrianBurnsSWL]

Hello Steve,

Thanks very much for the generous offer of help!

The mechanical end of things looks quite straight forward to me, as that is the sort of thing I've done for many years. Making jigs and fixtures for guitar production requires making metal parts as well as wood. I routinely cut, drill, and tap, cold rolled steel for instance, as well as make dust hoods for machines, and other miscellaneous parts out of sheet metal.

And yes, I would like to be able to go all the way up through 10 meters with the rig, though I imagine that lower power on the higher bands, due to a less than ideal final tank circuit, would not be a problem.

Wow! 33% efficiency for screen modulation, eh? That's right down there with cathode modulation! When I asked on AMforever about the lowest practical power for AM work, the consensus was 100 watts out. Maybe screen modulation would be the way to start.

As for 2500 V. B+, Wayne Mills N7NG still has a HUGE HV transformer that I scrounged from a diathermy machine back in 1957. I traded it to him for I don't remember what, and when he came by my shop a few months back, he said I could have it if I wanted it. He was using it for a door-stop! I never checked  the voltage out of it, as I was scared to hook it up--probably a good thing too!

Again, thanks a lot for the offer of help (:->)...


Cheers,

Brian

Cheers,

Brian

[KM1H]

Brian, read your PM's.

BTW, my Brian is heading to Djibouti for a 6 month deployment next week so let me know if there are any woods you might need.

Carl