Receiver multicoupler for HF

[W7TFO]

Does anyone have a favorite pet schematic for one of these, with say about 4 to 6 outputs?

Pre-war topology, no solid state.  No miniature tubes...this has got to be a period item

Regards.

[Pete, WA2CWA]

Some of the TV/cable 4-way splitters will go down to about 3 or 5 mHz.
Or you can buy some commercial products like this setup to connect 27 ham receivers to one antenna:
http://www.wy6k.com/homebrew_stuff/rx_antenna_distribution.htm
Or you can homebrew one:
http://www.dxing.info/equipment/rolling_your_own_bryant.dx

[Opcom]

I like the CU-168/FRR. It has 5 outputs and appx unity gain from 2-32MHz. The broadband RF coils/transformers are well documented in the manual (wire size and turns) and could be a challenge for homebrew at least to me anyway.

I use this and it is just great and it does not seem to add any noise. It also isolates the receivers very well because each channel has its own amp. The amps are balanced push pull. The input stage is grid driven and cathode-biased. The output is cathode coupled to the cathode of the next stage which is GG. The plates drive the OPT and the whole thing is very stable. It is 70 Ohms, but that has not been a problem for me and I have pretty much ignored it. If you have more than one of the couplers, at least two can be daisy-chained.

The input and output are co-ax, and a balun arrangement along with an artifical delay line is used on the input, and each amp has its own output transformer and RF jack (type N).

The amps can be removed and swapped around for troubleshooting.
Each amp uses four 12AX7's running on 130VDC.

Power is made from your choice of 105-115-125V input, with a 5U4 as the rectifier and the PSU is on its own sub-chassis too.

It is rackmounted and rather weighty. It is also very easy to work on and designed to run for years with no cares.

I keep the manual online, so to save disk space here (1.8 megs) pls download it there:
http://www.bunkerofdoom.com/mil/CU-168_coupler.pdf

There is also a CU-52/URR tube type coupler with 10 outputs that I got from Groban Supply, but it is a total basket case and I never got around to messing with it. There is a guy named Chuck Rippel who knows alot more about that model.

[Opcom]

Pete, those are the horrible solid state items! The man wants to be tubular! In any case, tube or SS, here is a schematic that includes a 37MHz low pass filter and a BC band filter, all 50 Ohms. It would be a good add-on for any coupler if the impedances can be controlled. Credits are Ku4AF

[W2XR]

I like the CU-168/FRR. It has 5 outputs and appx unity gain from 2-32MHz. The broadband RF coils/transformers are well documented in the manual (wire size and turns) and could be a challenge for homebrew at least to me anyway.

Hi Patrick,

I just took a quick look at the photo of this unit in the manual (for model CU-168/FRR) that you were kind enough to post.

Would you believe I saw one of these multicouplers getting ready to go into a dumpster this past June at a surplus joint that was clearing everything out? If I had known what it was at the time, I would have dived right in there for it; I have no pride when it comes to doing these kinds of things. It was in nice condx as I recall. I could have used it for connecting several of the receivers in the shack to a common receive port on my T/R relay. I'm currently doing this manually with a high isolation/low insertion loss rotary coaxial switch.

Live and learn.

I have seen a few of the earlier generation solid-state versions of receiver multicouplers at hamfests, etc.,  but I have always passed on them. I'm fearful there may be too many issues with spurs and IMD at high signal levels, but I could well be wrong on this.

TMC made a nice tube multicoupler; I don't recall their model no. or the military designation for this unit. It's probably fairly rare nowadays, like most of the TMC gear.

73,

Bruce

[WA1GFZ]

Pretty easy to build broadband splitters. I run a 1:4 all the time on 160 and 75. Do not need any gain to make up for the loss since the noise floor is so much higher than the RX MDS.

[WU2D]

Pre-War?

Well almost all receivers could use some help then...

I would start with a remote cutoff pentode of that era like a 6SK7 metal or a 6K7 as a band switched front end preselector with an RF gain control. Put an RFC in the plate and feed 6 separate Cathode Followers like 6C5's or 6J5 metal.

If that is not old buzzardly enough use a type 39 into Type 37's or 76's.

Mike WU2D

[k4kyv]

I wouldn't use any kind of active device between antenna and receiver.  It will add noise and reduce the dynamic range of the receiver.  Better to use a broadband transformer, or some kind of tuned circuit with a separate coupling link for each receiver.  That's what is wrong with those 50's era active electronic T/R switches.

Ideally, a receiver should be single conversion and completely passive from the antenna to the first i.f. filter, but it is hard to come up with a passive mixer that doesn't have too much loss to feed the i.f. amplifier directly with no r.f. amplifier stage.  But any coupling amplifier where the additional gain is not needed, SS or tube, will degrade receiver performance.  This includes outboard amplified pre-selectors.

If I want to run two receivers simultaneously off the same antenna, I usually just connect them in parallel to the antenna without any kind of splitter network, and I have never noticed any degradation other than maybe the loss of an S-unit or so of signal strength in each receiver compared to when the other one is disconnected.  I'm sure each receiver generated some birdies or garbage into the other one, but I never was able to notice it.

[W7TFO]

Well Don, that is just what I have done in the past out of expediency. 

It does generate a few artifacts here and there across the band, mostly in the form of a 'variable dead spot' on radio 'A' when tuning radio 'B'.

Simple is better in all cases.  I thought about doing this in terms of a broadband, untuned class 'A' RF amp with just enough gain to offset the loss of a dividing network.  That follower idea is a really good one.  Maybe a passive trap or two on the input to knock down the out-of-band flamethrowers out here.  I don't think that would raise the noise level to any extent noticeable.

I'd do this in a completely balanced design so as to be able to use my open wire system/dipole and my old Hammarlunds.  I'd probably benefit from CMRR this way as well.

Please, any more advice?

[WU2D]

I gave the cathode follower circuit a try. I put a simple bandswitched tuner up front.

It turns out that to get a low impedance output, you need a high gain, low plate resistance tube and for RF, a tube with low noise and high frequency gain. I immediately thought of our friend the 6J6 and putting both sections in parallel. The other trick is to separate the bias from the cathode resistance. This resistance is in parallel with the tubes output impedance, so to hit 52 ohms we must go low. This split also greatly boosts the apparent INPUT resistance above the actual grid resistors value of 220K in this case to 2.2 Meg.

How is that trick possible?

Mike WU2D

[WA1GFZ]

Mike, nice EMP isolator

[WU2D]

Frank,

All of those Racals, WJ's and such will burn up like a spectrum analyzer in a hillbilly radio shop.

Mike

[Opcom]

Mike, nice EMP isolator

EMP isolator would be a heavy planar triode in class A grounded grid, then follow it with the cathode follower coupler. The 2C39 might be an easy to get candidate.

The 7077 and 7486 were used in some front ends.

I also have to apologize for not paying attention to the requirement for no miniature tubes. The tubes shown in the schematic I uploaded were mini's. I think they have octal counterparts though. 6SC7, 6SN7, 6SL7, etc.

[Opcom]

I gave the cathode follower circuit a try. I put a simple bandswitched tuner up front.

It turns out that to get a low impedance output, you need a high gain, low plate resistance tube and for RF, a tube with low noise and high frequency gain. I immediately thought of our friend the 6J6 and putting both sections in parallel. The other trick is to separate the bias from the cathode resistance. This resistance is in parallel with the tubes output impedance, so to hit 52 ohms we must go low. This split also greatly boosts the apparent INPUT resistance above the actual grid resistors value of 220K in this case to 2.2 Meg.

How is that trick possible?

Mike WU2D

That's a puzzler. Perhaps since the gain is 0.9V/V, and (0.9*Vin) appears at the lower end of the 220K resistor (junction of the 220K and 120 Ohm), the resistor passes only (1/10 the normal signal current it otherwise would. The output signal voltage from the cathode, derived by the signal current through the 120 Ohm resistor, resists the input signal current flow in the 220K resistor. Does that make any sense?

[N2DTS]

I have some sort of splitter on 2 of my receivers, it was marked something like 1 to 100 mhz, and has one port marked DC block.
I cant tell any difference with it in line or not, no active components, all normal receivers have way more gain then they need on lower frequencies, as long as you use the transmitting antenna, most receivers seem to be designed to use short antenna's, the 80 meter dipole may be to much for them.

In my homebrew receivers, I dont have any RF amp, use a passive preselector, and I still have more gain than I need.

Brett

[WU2D]

I gave the cathode follower circuit a try. I put a simple bandswitched tuner up front.

It turns out that to get a low impedance output, you need a high gain, low plate resistance tube and for RF, a tube with low noise and high frequency gain. I immediately thought of our friend the 6J6 and putting both sections in parallel. The other trick is to separate the bias from the cathode resistance. This resistance is in parallel with the tubes output impedance, so to hit 52 ohms we must go low. This split also greatly boosts the apparent INPUT resistance above the actual grid resistors value of 220K in this case to 2.2 Meg.

How is that trick possible?

Mike WU2D

That's a puzzler. Perhaps since the gain is 0.9V/V, and (0.9*Vin) appears at the lower end of the 220K resistor (junction of the 220K and 120 Ohm), the resistor passes only (1/10 the normal signal current it otherwise would. The output signal voltage from the cathode, derived by the signal current through the 120 Ohm resistor, resists the input signal current flow in the 220K resistor. Does that make any sense?

Yes! This sounds stupid but the input resistance of the follower is simply the AC resistance that the generator is loaded by. By Ohms law it has the value e/ i in where i in is the current which flows through the grid resistance (220K). This current is equal to the grid to cathode voltage divided by the grid resistance (220K).

Let's assume the gain of the stage is 0.9. So for a signal voltage, of 1 volt, there is an output voltage of 0.9 volts (across the cathode resistor). The grid to cathode voltage across the 220K is the difference between 1 volt and 0.9 volts or 0.1 V. so the current is 0.1V/220K = 0.45 uA. Therefore the resistance that the input actually sees is not 220K, it is the signal voltage divided by the current or 1V/0.45 uA = 2.2 Meg.