[HBR] My approach to the HBR... Suggestions?

[Walt Hutchens]

I'm way behind on this discussion (due to press of work) but I'll toss a 
few comments from a few days ago.

Pete quoted me:
>> My favorite AGC circuit uses a plate detector with almost any of the
>> common triodes with the cathode returned to a negative bias of 50
>> volts or more <SNIPPED>

And commented:
> I redrew the schematic so I could get a better handle of what Henry was
> doing.  It looks too simple :)  Is the cathode voltage critical,or will any
> voltage over -50 or 60 volts work?  Also, is the original design, with VR102,
> 10K trimmer, for AGC Adj. needed in the HBR?  I don't see what purpose
> it serves, and it doesn't have much authority over the cathode voltage??

It's actually a Yaesu set, sold in a number of countries at the time;
if you peek at the S-Meter face it says FT-200 in one corner. I
believe it was sold under the Sommerkamp label in some European
countries. Henry Radio was the U.S. distributor. It's an excellent
vintage (vacuum tube with transistorized VFO) transceiver; superior to
anything else I can think of, for years in both directions. WAY better
than the FT-101, for example, although not so well-turned out, feature
wise or as rugged construction.

Maybe the KWM-2 ... and that's certainly better construction.

Anyhow, anything over -50 V or so would work for the cathode supply I
think but you will have to fiddle with the values of the various
resistors in the voltage dividers to get the AGC tube operating point
right. You can't go too high (negative voltage) because much of this
voltage appears between the heater and the cathode unless you use a
separate filament transformer.

Very little negative current is needed, so you can use a voltage
divider to cut it down. However the voltage should be well bypassed or
attack time will be slowed.

Actually, the closer you get to the original values and voltages, the
less you're going to have to fiddle, so I'd suggest trying to
duplicate the T-1 as a starting point.

The plate must be at 0 volts with no signal. The tube must be cut off
by just enough to establish AGC delay: basically when testing on a
signal generator (that is, with no external noise) you don't want the
S-meter to start to move until a signal is clearly audible over
receiver noise. One of the disadvantages of this circuit is the count
of small parts, plus it has a back-of-panel adjustment that others
don't need. However all the parts ARE necessary. ;-)

As I recall, there are two diodes; one of them is associated with the
use of the AGC when on transmit; that, of course, you don't need.  The
other one clamps the AGC tube plate so it can't go positive -- a very
good idea for an AGC driver stage.

That back-of-set adjustment sets the bias on the AGC detector tube
which sets the plate current and thus sets the plate voltage: you
adjust it so the plate is barely at 0 volts -- as I recall, it winds
up about -0.25 volts being almost as positive as it will go. This
setting the AGC delay value. Must be done with the AGC threshold
(front panel control) set to 0.

This circuit gives you a high-Z AGC line so any gas in a controlled
stage will cause trouble -- low S-meter readings and low gain,
possibly distortion, maybe a sort of squelch effect where the receiver
works okay on strong signals but goes dead when you tune away.  This
is a FREQUENT problem -- I'd say 50% of all tubes I've tried won't
work on a high-Z AGC, though you can use them for other things.

The advantage of high-Z is that you can get fast attack without having
to deliver much current or power. With triodes having a lower plate
resistance than the 12AX7 it would probably be okay to cut the two
plate resistances to ~1 meg and triple the cap setting the time
constant(s); that would somewhat ease the problem at the cost of
larger value caps.

>> You do need a tapped final IF transformer but the tap on the command
>> set output units works fine or you can use a capacitive tap.
> 
> I assume at this point you're talking about the IFT tap for the
> triode product detector?

Yes.

> I also further assume on the Command IFT I would be grounding
> the tap, and taking the signal from the lowest Z side of the winding?  That
> circuit also seems to use the CT to reverse the IF signal phase to the
> PD, so the neutralizing cap is able to cancel the BFO leakage?

Yes -- you need to reverse the phase to get the neutralization to
work.  I don't remember and can't easily check which side of the
winding I used for which.

>> The signal detector must be comparably high gain.  For SSB the usual
>> triode, gated or switched beam tube detectors work fine.  What I've
>> done, however, is use another plate detector, adding a BFO signal for
>> SSB.
> 
> Again, we're talking about the triode PD used in the Henry Tempo One?
> Are you happy with the PD?  The circuit looks deceivingly simple!
> And sure would make the project easier, I was going to use a 6JH8
> for the PD.

It works fine. One disadvantage is that the gain is different on SSB
than on AM -- for equal S-meter readings you have to turn the volume
up somewhat on AM. This seems to be built in: the BFO signal shifts
the tube operating point I think.

AFAIK about the only way to beat this is to provide separate signal
detectors for AM and SSB, with component values adjusted to give you
equal gain. In another receiver I've had sitting on the kitchen
counter in '50% done' state for a couple of years, I used the T-1
signal plate detector for AM but put in an 6BN6 product detector for
SSB.

This is the typical complexity/performance trade-off.

Use of a scope to tweak parts values is essential if your detectors
are to provide really good linearity. Feed in a sine wave modulated AM
signal and check that the AM detector delivers a sine wave, give the
SSB detector a carrier adjusted for a beat note and do the same for
that detector.

This might not be needed if exactly duplicating an existing design --
particularly W6TC's -- but I think many guys who designed the
build-it-yourself sets of the 60's were not very fussy about voice
signal quality.

(For that matter, RCA's beautiful SRR-13 (U.S. Navy, ca
1949) (by design!) barely copied AM and SSB was all but
unintelligible. Mechanical filters, a projection dial, nitrogen filled
tuning cap, temperature compensated coils, ... but the detector/audio
section is so non-linear that it's almost unusable for anything but
CW.)

Get a sine wave in to yield a sine wave out and it'll sound good.

Among the adjustments you might have to make is the signal level at
the output of the IF/input of the detectors so that on the strongest
possible signal the detectors still operate in the linear region.

>> One added complication is the need to neutralize the grid-plate
>> capacitance of the detector tube in order to prevent the BFO from
>> affecting the AGC. This isn't difficult: You build the circuit without
>> the cap, see what you get for an S-meter reading when the BFO is
>> switched on, and fiddle with cap values until the reading goes away.
>> The T-1 uses a 'gimmick,' I've used up to about 5 mmf mica.
> 
> I just want to make sure that the Tempo grounds the IFT tap to
> reverse the IF signal phasing to the PD, and that the PD IF tap is
> at a very low impedance point on that transformer...

That's correct if I understand the point you're making. Keep poking at
me, though, and I WILL get that receiver out and make the measurement.

>> With a 6EH7 RF stage, any of the better detector circuits (Pullen is
>> low noise -- not 'better' for an HF receiver having an RF stage) and
>> two stages of 6EH7 IF this will give you about the best that's
>> possible in an HF multiband receiver.
> 
> I may try the 6EH7, nothing is cast in stone, I see there are several
> semi remote pentodes that were made for TV IFs that look promising.

I think the 6GM6 is another very good performer.  But nothing I've
ever tried has the specs or the performance of the 6EH7.

One thing you have to look at is a rough match for the cutoff voltages
of all the AGC controlled stages if you use more than one tube type.
If one cuts off at -20 volts and another at -8 volts, then there's
going to be distortion on very strong signals because the one stage
will be doing all the gain reduction and will be right at cutoff.

>> If you use an RF stage, I don't think it's a good idea to use any gain
>> at the 1st IF. The 2nd mixer overload problems are considerably
>> increased by doing so.
> 
> That would be ONLY to overcome additional losses if I cascade IFs at 1600kc,
> and that IF stage would be throttled back to the minimum needed gain. It's
> easier to remove a stage than to add one... :)
> 
>> Three 6BJ6 IF stages at 85 kcs will give AMPLE
>> gain. I personally would go for two 6EH7 stages because of the amazing
>> overload tolerance (with high gain) of this tube, but using what you
>> have is a ham tradition and the three 6BJ6's will work fine.
> 
> And Peter has two sleeves of 6BJ6s, and nary a 6EH7...

You're talking to a guy who extracts many of his small receiver parts
from gone-dead computer monitor circuit boards and solders on longer
leads ...

>> With the AGC circuit described above you can even use sharp cutoff
>> tubes in the IF stages because the signal levels at the output are
>> never greater than the fraction of a volt range.
>>
>> Even matching the 85 kcs IFTs can make you fairly crazy if you don't
>> use something close to the original IF circuit because the tuning
>> range of is next to nothing.
>>
> I have two HFO rocks, one 1515kc, and the other is 1710.080kc. One or
> the other will hopefully work in this area.

If neither of these works out, put out a call; I have quite a few
assorted crystals and can probably find another one or two in this
general area.

No matter how you do it, there'll be a bit of cut-and-try to get the
IFs and the oscillators working together but this scheme will give you
a very satisfactory receiver.

I read the list regularly but can only reply via the webmail interface; replies by 
email are considered spam.  I have read the hints (inapplicable) and written to 
whyme but that address is not currently being checked I guess -- my note was 
eventually returned as undeliverable.  So I may be a bit irregular in 
posting.  

However it is inspiring to see the list active again!  Thanks!

Walt
KJ4KV

Walt
KJ4KV