HBR-influenced receiver

[KI4YAN]

Ok, I know it's been a year since y'all last heard from me. Maybe longer, I don't remember.

Been working on my car, not spent much time with the radio.

Anyway, I'm starting to start collecting parts for this year's "Winter Project" and so far, this is what I've dug up:

-2ea Miller 1732 1650Kc IF cans

-3ea BC453 85Kc IF cans

-LOTS of sockets, octal, 7 pin mini, 9 pin mini, UX5, UX4, etc

-2ea 3 section, 7-25pf tuning capacitors WITH slide rule dial mechanism

-A few old QST articles on the HBR series of radio receivers.

I've got other stuff like power transformers, chokes, etc running out my ears.

Has anyone here built an HBR-XX of any kind? Anything similar? (double conversion, plug-in coils)

Any suggestions on features to incorporate to make this receiver a usable, effective tool, in addition to fueling my toob-head addiction?

Anyone got any Xtals to trade in the 1530kC-1560Kc range for the second LO? (or, even better, one at 1565Kc?)

Thanks again for all the help guys.

[WB6NVH]

I am not very enthusiastic about 1650-1700 KHz IF's anymore now that the expanded AM broadcast band can swamp them, unless you want to add the effort to build up filters and wavetraps.  Depends upon how close you are to a station, whether it's day or night, and whether you want a 160 Meter band.

160 on my SX-101 was ruined by a foreign language broadcaster on 1650 but they disappeared before I had a chance to fiddle with filters.

[w3jn]

Indeed.  My HB RX uses a 1600 KHZ IF with xtal filters out of some Racal receiver.  My first mixer is a 7360 and my first attempt with the 7360 set up as an unbalanced mixer ended up with way too much BCB blast thru.  Re-doing the mixer in a balanced configuration, and carefully balancing the DC bias on the deflector plates pretty much solved this.  But it illustrates how critical good preselection is.  Mine was crummy - just one L/C element before the 7360 grid.

Those 1600 KHZ IF cans seem to have a pretty decent adjustability range.  See if you can't get 'em to adjust up above 1700 KHz, and calculate your 2nd LO xtal accordingly.  Doesn't matter what the xtal freq is exactly, so long as you can get the converted sig into the 85 KC transformers' bandpass.

I'm not a big fan of the HBR product detector circuit, nor the audio.  I suggest you change the product detector to a good dual-triode circuit (can be found in any older ARRL handbook), and pay careful attention to the audio section.  Use push-pull 6AQ5s, or a 6360 dual tetrode; phase splitter outta the RCA handbook.  This will make a world of difference.

[Ed/KB1HYS]

It might not take much to move the proposed IF up to say 2kc or so.
Get away from the BC band and 160 just about.

[WA5VGO]

I've built several of the Crosby HBR's.
-I've never had a leakage problem with the first IF, but you might want to consider moving it above 1710.
-The BC-453 IF transformers work very well.
-The HBR-11 works about as well as the HBR-16 with one major exception. If you go this route, definitely use the separate 6BH6 local oscillator. The triode of the 6U8 just doesn't have enough output on the higher frequencies since it works on the second harmonic.
-The hang agc is a major improvement, but I find it works much better when the voltage is derived from the IF rather the audio.
-I find the 6BY6 product detector to be quite good. On the other hand, the 6BE6 isn't so hot.
-Don't skim on the chassis. Off the shelf chassis are not rigid enough. It's got to be good and stiff.
-As designed, don't expect to use it on AM. It's just too narrow. On the other hand, it's a real pleasure to listen to quality SSB signals, and it's an acceptable performer on CW.

[WU2D]

Looks like this guy could set you up with the crystal that will put you above the BCB.
http://ve6sbs.sbszoo.com/

http://4sale.sbszoo.com/crystals.htm

Mike WU2D

[KM1H]

The HBR is a bit dated for performance but the basic concept was good in its day.

Id suggest a SB/HW series Heath first IF as xtals, LMO and filters are dirt cheap and everywhere. Configure as Heath did or do it completely different to get down to 85 kc. Use a 6GM6/6ES8 front end with a cathode follower. Keep early stage gain to just enough to overcome noise on 10M. Follow with a Lamb blanker and then get into filtering and gain. Nothing wrong with a 7360 either (except price) but they really need a constant load on the input which means a RF stage.

Even good Collins PTO's are showing up in the $20-40 range for ones generally not used in ham gear or a 390.
Here is one from a 51J4 listed a few days ago. http://tubes_tubes_tubes.tripod.com/id160.html

Carl
KM1H

[KI4YAN]

Obviously, if a radio is no good on AM, then there is going to be a problem. What can be done to widen it up some, if it's just too narrow? Can I run two stages at 1650Kc, then only one stage at 85Kc? that would make my passband about 5.5Kc, but might short me on gain. I could switch in second 85Kc IF stage for SSB/CW, which should bring my passband down to 3.6Kc. I could also stagger tune the 85Kc IFs, but that's a little advanced, and I don't really know how to do that.

As far as audio, I had planned a 12AU7 triode-concertina-6V6 PP output. three tubes, good sound, simple enough.

there are no BCB stations in my local area that cause an interferance problem, but I'll see what can be done to shift the first IF up a to 1710 or so. I did kinda want 160M, but we'll see what I get.

W3JN, you would not happen to have a schematic available would you? I'd love to get some more ideas on how to do things.

[WA5VGO]

There are lots of ways to increase the pass band.

The most obvious (at least to me) is to increase the frequency of the second IF. Instead of 85 KC, why not use 262 KC? 262 KC transformers are pretty easy to come by.

The coupling between the windings on the BC-453 transformers is adjustable. By pulling the pin up, you can over-couple the windings. That will increase the pass band to a very acceptable amount. The pitfalls with this are the gain of the receiver will go through the roof and the IF skirts will broaden way out.

There are plenty of other ways to do it. Some are easy and some are not so easy.

Darrell

[N2DTS]

I dont think you need double conversion to get a good AM receiver.
I built two single conversion 455 KHz IF no RF amp receivers and they seem to work better than anything else I ever used on AM, 160 to 40 meters anyway....

You should let the filter do the filtering, not the IF, and use a good filter.

If you was to ask me (and you didnt) multi conversion just adds noise and other problems.

Brett
N2DTS

 

[WU2D]

One thing about the older receiver designs is that you do not get all of your selectivity in the first IF. Most of these older designs distributed the selectivity or even did the majority of it in the final IF. So if you get a pass of 10 kc or so in the first IF (1750 or whatever) you can do the rest at 85 kc. So you can set your HBR to be an AM rig with a BFO or tighten the selectivity and make it a CW only radio. But you can not have both - or can you?

The Command Set 85 kc cans have the adjustable slugs which are excellent for setting up a 4 kc - 6 kc AM passband. You will be surprised at how nice this works.

For CW (or SSB) you pull them way out and you basically get the Q-5'r passband of 800 Hz. This works! I did it with my Q-5r for years. But it is a pain.

Of course this leads to the the ideal - to have two IF strips with independent AGC and selectivity, one optimized for AM and one set for SSB/CW.

This turns out to be a pain so the compromise for the designer is to do modest selectivity in the tunable IF's and control the selectivity in the front of the 455 kc IF with mechanical or crystal filters (the R390A). The remaining IF stages which constitute most of the gain remain wide. The disadvantage of this approach is the noise associated with the wide high gain stages and AGC issues. 

So can your HBR beat an R390A in every respect? It depends... On 80M or 40M where the 1750 kc IF is still viable for image rejection, and using distributed selectivity and better mixers, and equivalent stability oscillators, I would think so.
 
 
Mike WU2D

[KM1H]

Its easiest to do the final IF at 455 or 500 since thats were the parts avalability is high.

If you are a creative junk shopper then look for a dead pre SP-600 Super Pro and scoff the IF section. The first 2 stages are mechanically adjustable for continuous selectivity from around 16 kc down to around 250 cps. Or find a SX-28 parts whore and use their variable selectivity method.

For CW follow up with any of the zillion 455 kHz 8 pole crystal filters available and some gain equalization between modes.

The nice benefit of a real low IF or triple conversion is you can build LC selectivity. There is no extra noise generated in multi conversions if done correctly. The problem with xtal and mechanical filters is ringing and distortion. The problem with a single high IF is lousy skirt selectivity even with a crystal filter. There really is no easy way out to performance; the easy road is all compromises and then you might as well buy a Heathkit.

Carl
KM1H

[WQ9E]

Another option is building the set so that it is either triple conversion with a 455 2'nd IF and an 85 Khz. 3'rd IF or build both channels and select between them as the second conversion choice.  With both IF frequencies you could switch the AM detector between the 455 Khz. for broad and the 85 Khz. for narrow (selectable sideband AM) and have the product detector set up for the 85 Khz. IF.

I have my TMC GPR-90 coupled to a Hammarlund HC-10 sideband adapter which basically consists of the final IF, slot filter, and product detector circuitry from an HQ-170/180.  When conditions are good I use the audio from the GPR-90 and when things are rough I turn down the receiver audio and use the audio from the HC-10.

I had to add traps to several of my Hallicrafters receivers to take care of a station in Iowa that comes directly into the IF.  A simple external tuned trap at the antenna terminal took care of the problem for me.

Rodger WQ9E

[KI4YAN]

As for the 85Kc IF cans, if I have the rods pulled out, the passband of *just* the IF can is 5.6Khz. I would assume that with a 6AK5 driving it, the passband would be the same 5.6Khz, and would look something like the below:



Note, the scale is 1db/1kHz, and the image is from Kees Talen K5BCQ's website.

Given that this is a *usable* width passband for AM (It may not be, however) Would the passband narrow down if I used more than one stage of IF? i.e. , 6AK5-IF can-6AK5-IF can-7360?

If so, that's a problem. If not, and the passband would remain something about 5.5Khz, then I think that it would be perfectly fine, given the current band conditions at my QTH.

[w3jn]

Let's go back to some basic receiver theory.

The *only* reason for double conversion in the HBR design is for image rejection.  For a given front-end preselector, you'll get much better image rejection using a 1650 KHZ IF than a 455 or 85 KC IF.  This is more important on the higher bands than on 160 or 75.  You won't get much image rejection at all with a 85 KC IF.  So the 1650 KC IF provides the image rejection and a bit of IF gain, and all the selectivity comes from the 85 KC IF.

Next, for best dynamic range and cross-modulation performance the way to do it is to determine you ultimate selectivity immediately after the first mixer.  The reason for this is that very strong in-band signals can blast thru the front end and mixer, and overload the first couple of IF stages if you're trying to listen to a weak one and the AVC isn't reducing the IF gain.   Look at a L/C IF as having gradual degrees of selectivity as you move thru the stages, and the first stage will be subject to all kinds of crap getting thru the first IF xformer.

Most modern receivers do it this way, or use a 15 KHZ-wide roofing filter in the case of up-conversion/high-IF frequency schemes.  That said, I could cite any number of older receivers with more than acceptable performance that have their filters well downstream of the first mixer, or use multiple L/C IF stages to achieve their ultimate selectivity (the R-390 and R-390A are two such examples, each with different IF filtering strategies).

Do not be concerned about the lack of IF gain.  THere's plenty of gain available in *one* IF stage.  The only reason for mul,tiple stages is to obtain selectivity and a good shape factor, in addition to wider AVC action.  The more gain-controlled stages you have, the better the AVC range.  Figure on each stage with a 6BA6s has about 15-20 dB of AVC range.

Speaking of AVC, you can obtain lower distortion in your detected audio by using a separate AVC amplifier and detector, and use an infinite impedance (triode) detector for the audio (look in a 60's ARRL handbook for this circuit).  To do this you'll need an extra 6BA6 and IF xformer.  Tap off the plate of the last IF thru a 10 PF cap or so to the grid of a 6BA6.  The cathode should have a 500-ohm pot in series to ground so you can set the gainof this stage, and thus the AVC threshold/range.  Wire it up as a normal IF stage otherwise, with the grid going to AVC thru a 1 meg or so resistor, and the plate circuit going to another IF xformer.  You then detect the AVC using a 6AL5 or similar diode. 

The advantage of this is you can filter the AVC line to the bejesus to get rid of any AF riding on it, and in doing so, you don't increase the capacitive loading on the audio detector (because it's completely separate from the AVC detector), thereby limiting its tendency to clip at higher modulation levels.  Combined audio/AVC detectors are a tradeoff between loading of the last IF transformer (which produces audio clipping) and getting rid of any audio riding on the AVC line.  See Jay W1VD's receiver measurements for stark evidence of this - at lower AF frequencies he sees AF on the AVC, and high distortion levels at high modulation indexes in many receivers.

For the product detector, you tap off the secondary of the last IF (grid of the infinite impedance detector) with a 20 pf or so cap, to the grid of the product detector tube.  Now you don't hafta worry about BFO bleedthru to the AVC detector, or those stupid audio-derived AVC arrangements (I don't remember if the HBR uses those or not). 


My view:  if you're gonna go to the trouble of building a 16-tube homebrew RX, you have a blank canvas and can do it any way you want.  The advantage of the HBR design is that it's well documented and can be fairly easily duplicated.  However, the HBR design could be improved in many ways - Carl pointed out several, as did I.    And Rodger WQ9E's idea of using separate IF channels for SSB/CW and AM is outstanding.  For me, the long poles in the tent are oscillator stability and audio quality.  If those two factors are substandard, I'm never happy with a receiver  But that's just me.

As far as schematics go, get a '60s ARRL handbook.  The schematic for the 12AU7 product detector is in there.  An older RCA tube manual will give you a schematic of a complete hifi amp including the phase-splitter, etc.  If the schematic calls for 6V6s, you can sub 6AQ5s instead.  Don't worry about the tone controls, etc., unless you want to include bass/treble adjustments in your RX.

Good luck, and keep us posted on your progress!

[KI4YAN]

OK, after hitting the books, reading here, and talking to my buddy, an EE major who is currently studying the relevant topics in radio, this is what I've come up with:

I need a 6Khz Bandwidth.

My 1650Khz cans can provide 21Khz bandwidth at -6dB, and 50Khz at -20dB

My 85Khz cans can provide a 5.6Khz bandwidth at -6dB, and 14Khz at -20dB

If I retune the 85Khz cans to 100Khz, bandwidth at -6dB goes up to 6.5Khz, and at -20dB, 16khz. At 135Khz (use a 47pF cap instead of the original) it's 7.1kHz at -6dB, and 21Khz at -20dB.

I may use them at 85Khz to start, then if I need a wider passband, bump them up to 135Khz.

[KM1H]

You can make any LF frequency IF you want with toroids. Experiment with selectable coupling using C or L.

The dual channel approach was popular in homebrew around 1945-55 and pretty much went away due to complexity and cost. National and Hallicrafters both did good jobs with selectable selectivities at 80 and 50 kHz respectively. You can look at the NC-300, 303, and SX-88, 96, 100, 101 series, 115 and more for ideas. Now is the time to learn to work with ferrites; the choices and performance capability today far exceed the 50's.

John makes a good point about spreading out the AVC control over multiple stages. By taking the AVC signal off the next to last IF, amplifying, timing,  and splitting the levels you can have an overload proof radio and with no audio component to filter and upset the timing. The 6ES8 is the only variable mu dual triode Im aware of outside of a compactron and makes for a great AVC controlled mixer. The varying load can pull an oscillator so a Vackar VFO, LMO/PTO, or a xtal oscillator with buffer ensures stability. With a 6GM6/6ES8 up front and just enough AVC to keep things linear (read the full spec sheets with curves) and then let the IF stages handle the bulk of the AVC requirements. If you dont care about 20M and higher then there are other choices that do close to the above.

Once you take care of the front end, which includes the stability issue, then the rest is simply standard building blocks.

Ive always wondered how a 832A would be as an audio amp? I think almost everone has a few in their junkbox; it would certainly look different and take up little or no more room than PP 6V6's. At 250V it should last forever and run relatively cool.

Carl
KM1H

[KI4YAN]

Yes, the problem with 6ES8 is that it does not perform well at all between 10Mhz and 100Mhz. I never could get it to do a dang thing inside that range, but below 10Mhz, and above 100, it has few equals.

[K1ZJH]

The 6ES8 should be an excellent choice for a Pullen style mixer at HF.

One idea I've been mulling over is using a 1682 kHz 1st IF (have several HR-10 half-lattice xtal filters laying about) with the command receiver IFTs at 80 kHz for the second IF stage.  I've been wondering about using single conversion for AM with a 1682 IF stage, and then switching to dual conversion for the added selectivity on CW and SSB, or for "battle conditions" on AM.  Not sure it would be worth the work, but since it is a hobby it might be fun to try it.

Pete k1zjh

[KM1H]

The problem isnt with the 6ES8 and thats a given. What didnt the circuit do?   

The Pullen is the best mixer design and it really requires an isolated LO and it helps to have a cathode follower feeding a crystal filter if going that route.

I recently modified a HA6 with a 6ES8 Pullen and it is a major improvement to the point I dont even need the SS preamp any longer.

Carl
KM1H

[WU2D]

I do not know what magic that the BC-453 IF cans use but it is not "by the book" and you can get some serious selectivity with them. I assume that you are going with miniature tubes. Sure you don't want to go LOKTAL? Nobody fighting you for those.
Hey I put a Pullen in my Superpro with a 6SC7 audio tube right where the old mixer tube plugged in and it works slick.

Mike WU2D

[KI4YAN]

Argh. No matter how many times I measure, I never can get a chassis right the first time I bend it.

Aiming for 12"x15"x2", I made a wonderfully stiff, ridged 12"x12.5"x2" chassis.

I'll give it another go later.

[KI4YAN]

Ok, doing a little more thinking here, I pulled up an old mixer design that is normally used for VHF and up:

The cascode. It's low noise, and has decent gain. the availability of frame grid triodes allows for high gain and low distortion, without the added partition noise or spurs.

I am Switch, in this thread:

http://geek.scorpiorising.ca/GeeK_ZonE/index.php?topic=3515.0

Basically, I was fiddling around with television tuners, thinking that if it would work at VHF and UHF, then it could be applied at HF with good results. Greg reports that a 6ES8 cascode into a single IF can with NO rf amp and no IF amp was able to drive a 1n34A diode detector enough to pick up broadcast radio, although he did run an audio amp to drive his headphones.

Has anyone here had any experience with this type of mixer at HF? It worked FB on FM broadcast, when I built my 9-tube FM reciever, where the LO was running only 200khz from the desired signal. (Get much farther and images become a problem, but since FM broadcast channel spacing is about 200khz, I had zero problems with images. In fact, you could tune a little off the main channel, and null out nearby stations with it.)

[K1ZJH]

If  your dead set against using a Pullen, I'd suggest look at a gated beam mixer.. The 6AR8 is cheap, and has been used for mixers with good results.  Do a Google search for "6AR8 mixer" and you should get some hits.

Pete

[KM1H]

The 6AR8 is microphonic as hell and was quickly dumped by TV manufacturers. What you find on the market today is the endless recycling of over production leftovers.

Carl
KM1H