Multicoupler project, progress...

[WBear2GCR]

short version...
I bought the wrong multicoupler (that's an antenna distribution box, 1 in, N out).
The one I got was 30 - 300mc, not 3.0 - 30.0mc. Ooopsy.

But it was a very very nice box.
Came with a regulated B- supply, a very sexy amplifier strip, passive splitters, a filament supply (adjustable),
and well, you can see the pix.

So I tested the splitters, they're generic, total insertion loss - 26dB good down into HF, no problem.
The amp is super cool, but purpose built, not what I expected before it arrived. Uses a single WE planar
triode in GG input, and a row of 5842/417A tubes... thus the supplies.

I'm moving it to HF, to run coax around to various different receivers in different places...
AND, I'm embedding a DowKey TRM electronic TR switch so it can passively sit in the antenna coax feed
all the time, safely, with no other switching required (running up to ~1kw peak).

Some pix:

[WBear2GCR]

This will mount in the lab rack, and with the front and back BNC connections it will be possible to just patch in receivers as
needed, and have permanently fed receivers off the back side. The ANT will feedthrough front to back...

So what goes in place of the neat VHF amp strip will be the DowKey TR, a BC band filter, and a gain block.

Here's the original VHF amp strip: (a work of art!)

[WBear2GCR]

And then the gain block and the basic layout... note the regulated 150vdc power supply has slid over
and the constant voltage RAYTHEON ac transformer is gone.

FYI, anyone who is on FB can search for my call, there is now a private group, just click the join and reference amfone
when prompted...

[W1NB]

Out of curiosity, what is the NF on the amp?

[WBear2GCR]

I can not say, since I was unable to find a catalog for the company online!
But I expect it is commensurate with the better commercial amps of the time...
I have a Mini-Circuits module, that is smaller, those NF are published.

Regardless, it's going to be way less than the noise floor on any band I work, afaik.

What was your thinking on this?

                     _-_-

[W1NB]

I would agree. It looks to me as though there is a lot of silicon in there for 35 dB of gain. I’m wondering if there is some broadband AGC incorporated to reduce the chance of front end overload. It looks very well built.

[W1NB]

I don’t know if the following information will help you in finding more info about the amp. This comes from a lengthy summary of the M&A of the RF manufacturing business. I stumbled across it a year or so ago when I was trying to find out what happened to Microwave Associates.

“Q-bit was bought by Remec in 1997. Spectrum Controls bought the Remec Components business unit in October 2004. In February 2005, Spectrum Control bought Amplifonix.

Radian Technologies is part of Spectrum Microwave.

Remec the company no longer exists. It de-listed itself from NASDAQ and liquidated everything. But Remec the nameplate lives on as part of Chelton. In July 2005, Chelton sold the Outdoor Unit and Transceiver business in Poway, CA to Wireless Holdings International for $15 million. That same month, it also sold Electronic Manufacturing Services (EMS) in Escondido, CA to Veritek and Samjor. In May 2005, it sold its Space and Defense group in San Diego, CA and Richardson, TX to Chelton, a part of Cobham plc, for $260 million. In March 2005, it sold part of the Wireless group to Powerwave Technologies for $118 million. In May 2004, Remec sold its Fixed Wireless Access Systems (FWA) unit to Axxcelera Broadband Wireless, Inc. It sold its Antenna and Artificial Intelligence (AI) -related assets to the owner of Optimal RF, a startup.”

[W4AMV]

Q-bit was founded by one fellow I recall meeting, Dave Halloway(sp) who was with Radiation Labs... long time ago. The invention of the noiseless feedback amplifier by Dave Norton was improved upon by a technique that Q-bit devised; Power Feedback. Key to the design is a lumped element directional coupler. The looks of the inside photo, that appears to be what you have. The patent and circuit description are in this pdf:

https://patentimages.storage.googleapis.com/24/db/62/8da5679bb424da/US4879525.pdf

Keysight has circuit models for the Q-bit model line of amplifiers:

http://literature.cdn.keysight.com/litweb/pdf/ads2004a/vcsys/vcsys025.html

The claim to fame for power feedback compared to noiseless feedback is higher reverse isolation with all the benefits of noiseless feedback, low noise and high intercept or very low IMD.

Alan

[WBear2GCR]

WHOA!!!

Alan! How great is that!!
A million thanks!

I've got some stuff to read up on!!
Super kewl!!

             _-_-

EDIT: I had run into the Keysight models page. But don't know what I can possibly do with
their models, nothing I see to click on anyway. The model's listing does show some
specs... "Q-bit QB-538, 2.0-500.0 MHz, Po1dBC=22.00dBm, NF=3.0dB, Model features SP, NF, 1dBC, IP3"

[WBear2GCR]

A little bit, right-in-front-of-my-nose, type work shows a patent number ON the case!

It's a more basic patent than the one Alan put up, which is more or less the exact circuit
in the amplifier. This patent shows the circuit engineering more specifically.

[WBear2GCR]

Out of curiosity, what is the NF on the amp?

Per the above posts:

NF=3db

[W4AMV]

Sure... And it get better than that! Ah you beat me to it! Yes, you got the patent no.

On your amp you have the patent no.

4042887. And if you pull that patent up you find the names of the founders... at least one for Q-Bit, that I recognize. And I apologize, I was close. It was Gary Callaway.

https://patents.google.com/patent/US4042887A/en

So this is the exact description of what you have and again a Power Feedback Amplifier Configuration with the unique arrangement of the directional coupler.

I should add, much of the Silicon in there is not part of the signal chain, but used for
active bias. The RF device emitter is at DC ground and conventional R bias is not used. Instead another transistor is used to provide DC negative feedback by sensing the collector of the RF gain stage and subsequently controlling its base current.

Alan

[Opcom]

The planar triode was possibly there to protect against electrical mishaps at the multicoupler input. You did keep the amplifier strip, right?

[WBear2GCR]

Yes, Partick, I have it, but do not forsee a use.
Nor can I locate a schematic!

I'd keep the 5842/417a, even though they are Amperex.
The planar tube is neat, but again, can't see a use now.

Happy to pass it along!


Alan, this is extremely cool stuff. Only last year I downloaded an article on the Norton amp
design, which somehow eluded finding me for apparently decades! Shame, since that is very
cool. I need to print them out and look at that R sense set up! Of course I'm what, 2 decades
plus late to this party too!! Oh well...

Apparently only you and I find this interesting?

[WBear2GCR]

Just did this preliminary layout for the 24vdc PS to run the Q-Bit amp...
Will have a two stage Vreg, coarse, and then final vreg out... decided to go
fully shielded, and keeping with the overall design, send the power out on
coax... the small parts will go on some solderable protoboard cut to fit in the
center where the caps are sitting now.

The Bud box PS module will just shoehorn into the open space next to the
black 150vdc regulated PS...

Btw, just got some Chicom connectors and adapters in from... Wuhan!!

                  _-_-

[Tom WA3KLR]

Bear,

Below is the link to a 1970's R & S multi-coupler I spotted on the web a few years ago. It is the complete manual and schematics; a reference for some ideas perhaps.

https://www.deimert.se/nv14t/nv14t.pdf

[WBear2GCR]

Thanks Tom,

Yeah, this R&S is fine, but very parts and labor intensive...

there are two main methods; active & passive.
The R&S is what I would call passive. It uses a preamp to make up the insertion loss
of a multiway splitter. One in, N out.

That's what this one is. But I am using the Q-Bit amp section which is super high performance,
and all in that machined box. Today they use MMICs, which may or may not equal or exceed the
Q-Bit unit.

The other method is all active. That uses a main amp, which can be a buffer or low gain, running into
N buffer amps that drive each recvr out.

I have a TMC example of that. Unfortunately mine has only
two output boards, and it will take up to 8, each one with two buffers, so 16 out total! I could make
up new boards, but I'd have to find a way to make or buy or cut off that size slide-in finger contacts
to mate with the connectors... slide in boards. Or go through some other process to hardwire, etc...
and it becomes a very long and complicated process...

Think this is the correct pdf...

[Tom WA3KLR]

Bear,

Interesting comparison of the two designs – The R&S uses one totem-pole emitter follower driving all 10 outputs with series 50 Ohm resistors whereas the TMC coupler uses the totem-pole follower driver driving 16 individual single-ended emitter followers with 47 Ohm output resistors.

Off the top of my head the 2N3866 and what R&S used were the low noise figure high current bipolar amps used in the CATV stuff.

They all were interested in very low intermod/high intercept point circuitry, but I wonder if you could get away with the puny microwave bipolar low-noise lower current transistors (Can’t recall the part number off the top of my head I am thinking of).

I like the individual emitter follower for each output like the TMC rather than R&S.  Those multi-couplers presumed everything connected would be a receiver-only.  I’m thinking in your application you may have a rice box connected at some point which accidently goes into TX.  I would try to get away with one emitter follower driving 6 – 8 emitter followers, all single-ended stages.

I am not familiar with the Q-bit amp, perhaps like the R&S you just have series resistors and caps from that amplifier's output to all output jacks.  This is probably what you are thinking.

The application is just a number of your own SW receivers running?

Anyhow I always see the big part of a project is the basic cabinetry – a.c. cord, fuse, power switch, power supply, and the coax jacks and in this case and you have it all there to start with.  

Enough already,
Good luck OM.

[WBear2GCR]

Tom, the splitters are commercial devices, sealed box. 50ohm in, 50ohm out...
If I transmitted into it, that would be "bad".

In essence what R&S fabricated.

Mostly only receivers.
Maybe I will make up a heavy duty back to back diode module to go in series
with any xcvrs that get plugged in... to go outside at the BNCs on the panel.
good thought there!

The interesting thing is that there is always distortion added with an active circuit,
whereas the passive one drops signal level, and only has the intrinsic noise and
distortion of the resistors themselves... but actually I like the active out approach
too, but finding a simple way to build up the output modules for the TMC would be
a lot more time and effort than this build... but it's still on the table, "later".

[Tom WA3KLR]

Since the multi-coupler amp is broadband, you want to have a hi-pass filter above the broadcast band and a low-pass filter below the FM band or TV Ch 2 unless you are far from any BC, FM, or TV broadcast transmitters.

R&S input protection is a spark gap and rf choke to ground and bi-directional diode clamp referenced to 11.2 Vdc.  The TMC has a neon bulb and their diode clamp is 3 diodes in series, both polarities.

One other interesting thing about the R&S schematic is their supply bus filtering in the rf circuit areas – they have 1 uF or 2.2 uF caps with 1.2 Ohms in series to ground.  I have never seen this done anywhere else. If you think about the normal practice of putting a cap to ground from a supply bus to filter out rf or other noise ripple, the ripple/rf voltage is converted to a current running in the circuit and not dissipated.  So my guess is that the 1.2 Ohms is to dissipate the undesired a.c.; a more effective bypass? It’s always interesting to study other people’s designs!

[AG5UM]

I've been watching this post with interest,
but I did'nt want to bother you with explaining the circuits to me, while your working on it.
So, I did'nt ask, So, these posts discussing the function and circuitry are very interesting, good post,interesting topic..
Now if I came across one of these I'd have a better idea of what it can do..
Cool
AG5UM

[WBear2GCR]

Fwiw, Mike, KE0ZU has put up a DIY version of this here.
Simple enough, you just need the 1 in: N out splitter, gain to make up the insertion loss of that
splitter, and the supporting circuitry and hardware...

see attached. his post has pics of what he built...

[W4AMV]

In the process of looking at some prior QEX papers, I ran across this one by Zack Lau.

Worth running and an experiment to duplicate and then drill into refinement of the directional coupler feedback amplifier.

73' Alan

[WBear2GCR]

Essentially, this is what is inside the Q-Bit amplifiers. They were a patented circuit. The use of the "directional coupler" for
the feedback improves several parameters over prior art. Tnx for the info!

(btw, all of this is new to me as of a few weeks back...)

               _-_-

[Opcom]

Yes, Partick, I have it, but do not forsee a use.
Nor can I locate a schematic!

I'd keep the 5842/417a, even though they are Amperex.
The planar tube is neat, but again, can't see a use now.

Happy to pass it along!


Alan, this is extremely cool stuff. Only last year I downloaded an article on the Norton amp
design, which somehow eluded finding me for apparently decades! Shame, since that is very
cool. I need to print them out and look at that R sense set up! Of course I'm what, 2 decades
plus late to this party too!! Oh well...

Apparently only you and I find this interesting?

Maybe so. Takes some special quality to appreciate the finest of certain stuff, I guess!

I wonder if that line of tubes there the original strip amp is a distributed amplifier, considering the bandwidth and the number of stages. I mean, 10dB VHF television amps only had a couple of tube stages in series.

Most distributed amps are wider in bandwidth than where they're used. It might surprise you on the low end.
This example, h/p 460, works from 3KHz to 140Mhz, gives 20dB gain.
http://hparchive.com/Manuals/HP-460AR-SN-046.pdf

Sure if you don't want it, but if you have time, see if the anodes are all connected at taps on an artificial transmission line, and same for the grids on a different transmission line.  -like in the h/p manual. - the lines are the little resistor-looking chokes in that h/p, and maybe in that strip too. h/p drew them tapped, but in reality they are not - each physical one is the end of one and the beginning of the next.

wishful optimism on that.. distributed amps are a favorite thing.