IF bandwidth for AM reception?

[KI4YAN]

What would an optimum IF passband be for AM reception? I have a set of BC-453 IF cans here (85Khz), including the BFO coil, and am planning to build an HBR-16 receiver.

These IF cans are reported to have a passband of 2.9Khz when the micarta rod is pulled out, and 5.3khz when the micarta rod is fully pushed in.

Is 5.3khz a wide enough passband for AM, but still narrow enough for SSB?

[WBear2GCR]

Yes and no.

A single IF can will not usually have enough "Q" to make the skirts very narrow. So that means that while the -3dB point will be where it is spec'd, the stuff out at say 10kHz away will not be down sufficiently - that means that if you are on 3880 and someone with a +40dB signal comes on at 3890, they'll be just as loud as the signal you are trying to listen to...

Selectivity in an old type receiver that had no "filters" came from cascading several stages of IF, with each stage adding to the selectivity. Even at that, it often did not provide sufficient selectivity.

That is/was the impetus for various filter methods, including the mechanical filter and the crystal filter.  The R-390 uses a tuning method to optimize the "Q" and so the selectivity of the receiver, iirc.

You say you have a set of filters, they might provide a wide enough bandpass for listening to AM, again depending on the final shape of the passband. You can likely "detune" them WRT to "Q" and get a wider passband too... at 85kHz you can easily implement a passive LC bandpass filter, or a modern IC based filter (various circuit and chip configurations are available).

Just my quick thoughts...

              _-_-bear

[W3RSW]

About 8 or 9 kHz, given a good shape factor, is a good compromise for most AM work.

Sad to say I used to think the 6kHz AM filter on my Drake was restrictive.
But these days with limited HF hearing, 6kHz sounds fine. 
On my QS1R I've ramped filters up and down in width on good BC stations and found that above 2.7Khz or so I'm not hearing the highs.   So I guess your filter pick might be based on your hearing.

[k4kyv]

From my experience with Collins mechanical filters, 4 kHz is normally too narrow for AM, but can be very useful under crowed conditions, and beats using a 3.1 kHz slopbucket filter for AM.  8 kHz is often too wide under normal amateur band conditions.  6 kHz is more optimum for general AM reception, but when condx permit, 8 kHz sounds much better.  I also have a 9.7 kHz filter, but it appears to be identical to the 8 kHz one.  I suspect the only difference is the nominal bandwidth stamped on the filter.  I can tell only a slight difference between 8 kHz and 16 kHz on the 75A-4, but on the R-390A the 16 kHz definitely sounds better when copying a broadcast station with good audio (something becoming more and more rare).

[KI4YAN]

Well, I'm going to be using these IF cans regardless, but I can modify them by trimming plates from the capacitors to bring the frequency up some. May have to remove some turns, from the coils, though. I guess I'll need to stagger tune the 2nd IF strip to get enough bandwidth.

First IF is likely to be a 1425Kc transformer from another command set, or I'll wind it myself. So far, I haven't found one. Idea is like so:

 RF -> 1st converter -> 1425kc -> 2nd converter -> 85kc -> IF amp -> 85kc -> IF amp -> 85kc -> infinite impedance detector -> 1st audio

I'm seriously considering an all triode receiver, for lowest possible internal noise. Yep, it's true, due to partition noise, the more grids you got, the more noise you got. Cascoded triodes are less noisy, and have just as much gain as a pentode. Plus, I have LOADS of nine-pin sockets.

It'd go something like this:

6BQ7 cascode RF amp -> 6BQ7 cascode mixer -> 1st IF can -> 6BQ7 cascode mixer -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 12AV6 infinite impedance detector -> 12AV6 AGC diode detector ->  12AU7 1st audio, 12AU7 phase splitter, triode wired 12L6 Push-Pull output.

Or, I could just go with the pentodes...

[ab3al]

why worry about it just buy a sound card and download some sdr software

[KI4YAN]

Why bother with AM at all then? If you are going to use a computer, why not just send an email?

Pardon my snap remark, but it's true. Part of the reason that I build my equipment.

[k4kyv]

6BQ7 cascode RF amp -> 6BQ7 cascode mixer -> 1st IF can -> 6BQ7 cascode mixer -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 12AV6 infinite impedance detector -> 12AV6 AGC diode detector ->  12AU7 1st audio, 12AU7 phase splitter, triode wired 12L6 Push-Pull output. 

And no BFO.  Then you would have justification for ignoring slopbuckets!

[w3jn]

Good plan but I'm wondering why you're using 6DJ8 cascode amplifiersin the IF?  No reason to use low-noise tubes there.  Better choice would be 6EH7 frame grid semi-remote cutoff pentodes, or even 6BA6s.  Better AVC control.

[N3DRB The Derb]

where's teh nu-vistors in all this action?

[K1ZJH]

Well, I'm going to be using these IF cans regardless, but I can modify them by trimming plates from the capacitors to bring the frequency up some. May have to remove some turns, from the coils, though. I guess I'll need to stagger tune the 2nd IF strip to get enough bandwidth.

First IF is likely to be a 1425Kc transformer from another command set, or I'll wind it myself. So far, I haven't found one. Idea is like so:
 
I'm seriously considering an all triode receiver, for lowest possible internal noise. Yep, it's true, due to partition noise, the more grids you got, the more noise you got. Cascoded triodes are less noisy, and have just as much gain as a pentode. Plus, I have LOADS of nine-pin sockets.

It'd go something like this:

6BQ7 cascode RF amp -> 6BQ7 cascode mixer -> 1st IF can -> 6BQ7 cascode mixer -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 6DJ8 cascode -> 2nd IF can -> 12AV6 infinite impedance detector -> 12AV6 AGC diode detector ->  12AU7 1st audio, 12AU7 phase splitter, triode wired 12L6 Push-Pull output.

Or, I could just go with the pentodes...

The HBR website has data for modifying the 1425 kHz transformers to cover in the 1600 kHz range; it is easily done by replacing the fixed silver mica button caps with 120 to 140 pF SMs.  That will keep the mixing scheme in line with the original W6TC HBR designs; I'd judiciously pick an IF that ends in "5" someplace between 1605 and 1625 that is selected to avoid any strong AM signals.

You will find that the 85 kHz IFTs have an extremely narrow tuning range, the trimmers are under 20pF.
They basically serve to correct for small wiring variations between different sets.  You could try sweep tuning the 85 kHz IF strip for a wider AM response, but my fear is that some of the trimmers will not have enough authority to do so. I'd stick the with the 6BJ6s specified in the original design, the first RFA and mixer will pretty much determine the receiver NF. The suggestion for using 6EH7s in the IF stages has some merit, but they have a lot of gain, and you'll need to watch the layout and stage isolation pretty closely if you use them.

Pete

[Steve - WB3HUZ]

I'm seriously considering an all triode receiver, for lowest possible internal noise.

As JN said, not really necessary. Your noise figure will be almost exclusively determined by the receiver front-end, the RF amp (if you have one) and the mixer.

[N3DRB The Derb]

is there a tube that has less noise for a RF front end than the nuvistor?

[K1ZJH]

One other suggestion:  if you're going for a strictly AM only receiver, why not just go single conversion with a higher IF (1625 kHz or so?)  Or, double conversion with an IF at 455 kHz?  Wider AM bandwidths using LC components is difficult due to the effects of arithmetic selectivity.

Pete

[KI4YAN]

The all triode reciever was more of a stunt than anything but the low noise figure was the only merit I think it had. Anyway, here's the more realistic layout:

6DJ8 RF cascode -> 6DJ8 Pullen mixer -> 1425khz 1st IF transformer -> 6BL8 mixer -> 85Khz 2nd IF transformer -> 6BA6 -> 85Khz 2nd IF transformer -> 6BA6 -> 85khz 2nd IF transformer -> 12AU7 Pullen detector -> 12AT7 1st audio -> 12AU7 phase splitter -> 12L6 push pull output. BFO will be fed into the last IF transformer, i think.

Yes, the nuvistor front end can be bested with a 6DJ8, or other low noise triodes. I wish i had a pair of nuvistor sockets though, the 2EG4 makes a *good* cascode front end and I have a pair...

As for being strictly AM, this reciever will not be. Needs to be able to cover AM, CW, and SSB, I just omitted the BFO earlier as it was a needless complication to the All triode theoretical receiver.

[w3jn]

One suggestion:  use a 6ES8 variable-mu dual triode in yer front end cascode amp.  The AGC action will be much better.  The 6ES8 is a frame grid dual triode designed for TV tooners - those needed really good low noise performance while maintaining good strong signal handling capability.  You can get 'em for a couple bux at a hamfest, nothing really uses them anymore and the audiophools don't really like 'em because of the variable-mu characteristic (which is desireable for variable gain circuits!).

You should ALWAYS use AGC on yer RF amp.  Running it full scrote is unneccessary, creates IMD products, unnecessary noise, overloads the mixer, etc.

ANother interesting thing might be to use a 6360 (9 pin min, dual power tetrode) for yer shove-yank audio amp.  Believe the one toob is capable of 8-10 watts by itself.  There was a circuit some years ago modifying the R-390A audio deck using this toob.

[Rob K2CU]

The low frequency IF transformers usually have fairly high Q by themselves. How you drive them and coulple them can give you pretty acceptable performance for what you are looking to do. I suggest you look at the coupling arraingement in the hallicrafters radios, like the SX-117. The 50 KHz transformers were driven by high impedance tetrode and could give pretty good selectivity curves due the the individual high Q of the transformers. the different bandwdths were achieved by changing the coupling betwen two successive transformers. You will need to use Silvered Mica or other very low loss caps to get similar results. Those transformers are loaded with ferrite sleeves over the coils besides internal slugs for tuning. Layout and attention to RF ground loops is essential too. Bu tyou have to drive them with a high impednace source, such as a tetrode/pentode. A triode just will not do.

[W3RSW]

Motorola used to love 6AK5's
It was "the darling of RF amps everywhere" in the 60's.

As far as wide band video goes, the 12BY7 is nice.

Well, better quit or another couple of lines will yield all I know.

[WU2D]

I agree with the 6BA6's in the IF (actually the first IF too) and you will like the 85 kc cans from the command set. I would adjust the slugs for max AM bandwidth and put a Q multiplier on the first IF (1425 kc). This will allow the HBR to do AM, SSB and CW bandwidths with control. The Q multiplier need only be a triode section because you will want to peak the iF, I imagine.   

Mike WU2D

[K9ACT]

Why bother with AM at all then? If you are going to use a computer, why not just send an email?

Pardon my snap remark, but it's true. Part of the reason that I build my equipment.

Snap is right. 

A sound card and software does not make a Software Defined Radio.  Many of us have built SDR's and get the same satisfaction as from building a boat anchor.

The simple practical facts are that, while you can build a tall ship that sounds as good or better than anything you can buy, you can not build a receiver with the selectivity and flexibility of SDR unless you use a computer.

My S40B with SDR IF is in another universe compared to what can be done with IF cans and traditional filters.

Time to expand your horizons.

js

[flintstone mop]

Hi All
Tim I thought Nuvistors were better suited for VHF and UHF.
The noise floor in the MF is pretty high from Ma Nature and man-made stuff.
And sometimes too much RF amp can be a disaster in the receiver, as local BC stations, or other close-by hot RF signals, will drive the front-end into overload and images.
Phred

[N3DRB The Derb]

they are better suited for that, but I just assumed we were talking theoreticals. atmo noise and such would make them unnessecary until you hit 10 M, and even after that it would be iffy. I have a AMECO preamp that uses them and below 28mhz it's worthless due to noise, but I tested it with Johnny Novices's 75 A-2 and it makes it a pretty hot reciever on 10 M. as compared to the naked front end.

[KM1H]

With the 6ES8 Pullen there is no need for a cascade RF stage unless you plan to use it on 2M.

All you need is a 6GM6 RF amp which will be nice and stable on any band and result in a system NF well below HF galactic noise on 10M.

http://tubedata.tigahost.com/tubedata/sheets/135/6/6GM6.pdf

Carl
KM1H