[R-390] More about the SSB prod detector

Fri Jul 29 03:06:04 EDT 2005

Hello again Dan:

Finally got some time (it's 2 AM and I couldn't sleep) to look over my 
product detector circuit, and you're (almost) 100-percent correct about the 
circuit details. Gosh, I'm embarrassed, and a bit nonplussed, when I think 
about the possible souls out there who may have flushed the circuit down the 
proverbial tubes when the AGC amp didn't work due to my drawing errors! Did 
that back in the year 2000.

In any case, I have already corrected my schematic print, and hope to scan 
it soon and update the Web posting. I also plan on crediting you with sharp 
eyes and much patience, if that's okay with you.

My actual circuit certainly does have a gain-boosting source bypass cap 
across the 1-kohm source resistor in the FET pre-amp stage. However, as I 
indicated in my previous note, the rule of thumb I follow is that the 
reactance of the capacitor should be about a tenth of the source resistor's 
value (or cathode resistor in the case of a tube, or the emitter resistor in 
the case of a common-emitter amplifier using a bipolar transistor).

The actual cap in my circuit is a 0.01-uF Mylar type. Using the formula for 
reactance Xc = 1/2*pi*f*C yields a reactance of 34.9-ohms. While that's not 
exactly 1/10th, it's close enough for government work, using standard parts 
from the junkbox. Using a 0.1-uF cap, as you note, would drop the reactance 
at 456-kHz to 3.49-ohms, which seems far too low. The approximate 35-ohm 
value is close. A 0.001 would providde 349-ohms (likely too high).

As for the polarity error on the follower stage, I can only wonder why I 
didn't spot that as I drew and scanned the diagram. That PNP stage could 
have just as easily been implemented with an NPN device, and then I probably 
wouldn't have made that error.

My cap is indeed inserted with the plus end at the highest point of 
potential, which (obviously, now) is the Vcc line. It's purpose is to ensure 
that the emitter is close to ground impedance (low). It probably ideally 
should be connected from the emitter of the PNP follower stage to ground, 
and then the polarity would be okay as shown, if it were drawn that way. But 
that's not the way I hooked it up. In any case, you want the PNP stage's 
emitter to have no 456-kHz energy there, thus the bypass's function. You can 
poke your scope probe there to verify that.

Please feel free to punt this e-mail over to Ed if you wish to.

My sincere apology for your inconvenience!

Regards,

      AI2Q, Alex    .-.-.

----- Original Message ----- 
From: Dan Merz
To: 'AI2Q'
Sent: Wednesday, July 27, 2005 8:35 PM
Subject: RE: SSB prod det. Hartley

Alex,  thanks for reply.  You might want to fix the schematic in case 
someone else wants to use it.
The polarity is wrong on the cap at the emitter connection on the pnp.
A bypass cap ( 0.1 mfd) on the source resistor of the 2n3819 was omitted in 
your schematic.

I've been in communication with Ed Newman and got a tip on reducing the 1000 
ohm in the source circuit of the 2n3819 to get more agc voltage out if 
needed.

best regards,  Dan.

From: AI2Q [mailto:ai2q at adelphia.net]
Sent: Tuesday, July 26, 2005 6:42 PM
To: Dan Merz
Subject: Re: SSB prod det. Hartley

Hi Dan:

Gosh, i haven't looked at that circuit in such a long time! The cap's 
polarity as shown is correct, if I recall.

You might also want to try raising the FET's drain resistance in order to 
see more signal swing. Do you have a scope? If so, have you measured the 
level of the IF signal coming in from the R-390A? Is it being amplified by 
the FET? These devices are quite susceptible to failure, especially from ESD 
during handling. Is the FET stage working? Checking that would be my first 
point of test Dan.

== Alex

----- Original Message ----- 
From: Dan Merz
To: 'AI2Q'
Sent: Tuesday, July 26, 2005 5:54 PM
Subject: RE: SSB prod det. Hartley

Hi Alex,  I mostly have the product detector for the 390 working after 
bypassing the middle oscillator buffer amp - the Hartley voltage was big 
enough without the middle stage.  The PD works ok.  But I'm having trouble 
getting the agc to work.  I noticed in your website diagram,  the 
electrolytic 4.7 mfd cap in the pnp transistor emitter circuit following the 
FET in shown with the negative side hooked to the positive supply voltage. 
Is this the correct polarity on the cap?  I decided it wasn't but now with 
no output on my agc circuit,  I'm wondering if I should have hooked it the 
way you show it and this is a "trick" to get lower cap impedance for the 
positive current pulses that contribute to developing agc voltage.  Ed 
earlier suggested going to lower source resistance in the FET circuit ( drop 
1000 ohms to maybe 100 ohms) to get more output but so far with 1000 ohms I 
get nothing out.  Was I wrong in reversing the polarity of the cap compared 
to your diagram? Or was there an error in the diagram?  best regards,  Dan.

I sent this once without changing the title - a long ago 390 thread - and 
decided it needed a better subject title.

From: AI2Q [mailto:ai2q at adelphia.net]
Sent: Wednesday, June 29, 2005 5:48 AM
To: Dan Merz
Subject: Re: [R-390] re cost reduction ect ect

Go with your textbook circuit Dan. It sounds like a Hartley, based on your 
choice of a tapped coil. In that case, I think you're right; tap the output 
at the drain load resistor. It should work just fine.

Hint: build your circuits on small modules so that you can re-vamp them as 
needed. If you build everything on one board, and the oscillator, for 
example, doesn't work as expected, you don't have to worry about sacrificing 
all the other stuff on that board when you scrap your first oscillator 
design. Take a look at my KISS SSB rig to get an idea about what I'm talking 
about.

QSL?

Vy 73, AI2Q, Alex  .-.-.

----- Original Message ----- 
From: Dan Merz
To: 'AI2Q'
Sent: Wednesday, June 29, 2005 1:50 AM
Subject: RE: [R-390] re cost reduction ect ect

Alex,  thanks for reply.  Just to be sure we're talking about the same 
thing.... I'm wondering about the output from the left-most mpf102 as a 
free-running oscillator to the 2n2222,  the next stage after the oscillator. 
I understand the 1496 needs low level input from the right-most mpf102 that 
the 2n2222 feeds,  which is what I think you're talking about.   If I change 
the xtal controlled first mpf102 to a tapped coil oscillator in the 
gate/source,  then should I feed the following 2n2222 from the source or 
from the drain of the oscillator.   My first inclination  was to feed the 
2n2222 from the drain of the oscillator like you did with the crystal 
oscillator but when I started looking at most handbook circuits the output 
of  free running oscillators to the base of transistors like the 2n2222 is 
almost always shown  from the source connection and not from the drain 
connection.   I would still be interested in seeing Ed wb2lhi 's free 
running oscillator circuit which I presume he used in front of a buffer 
circuit/FET follower similar to your circuit.  I'm at best a novice at ss 
circuits and fight my way thru understanding/copying circuits of others. 
It's all a marvelous adventure,     best regards,  Dan.

From: AI2Q [mailto:ai2q at adelphia.net]
Sent: Tuesday, June 28, 2005 7:36 PM
To: Dan Merz
Subject: Re: [R-390] re cost reduction ect ect

Hi Dan:

The FET source-follower will deliver a low-Z (perhaps 1-kohm) output to feed 
the mixer chip, at a gain of less than unity. If you use the FET 
reconfigured as an  amplifier you'll pick the output off the drain, but why 
would you want a higher level? The 1496 wants a low level (check the 
datasheet).

Vy 73, AI2Q, Alex .-.-.

----- Original Message ----- 
From: Dan Merz
To: 'AI2Q'
Sent: Tuesday, June 28, 2005 4:51 PM
Subject: RE: [R-390] re cost reduction ect ect

Alex,  I finally got around to building your outboard product detector.
I've got your schematics but I'm going to use a free-running bfo,  probably
in a circuit using an old NC 100 bfo tapped coil/cap assembly but I'm trying
to marry it to your circuit.  Do you have a copy of Ed's bfo circuit that
you could pass on to me to see how he did it.  I have an MPF 102 that I want
to use and am pondering whether the output should be taken off the drain or
source.  If I read your schematic correctly with the two xtals,  the signal
comes off the drain.  Most free-running oscillator circuits I find seem to
have the signal taken off the source side.  Perhaps it doesn't matter as
long as the right coupling impedances are in the circuit?  I'm matching your
circuit once I get out of the oscillator into the 2n2222.  Any help
appreciated,  Dan.

-----Original Message-----
From: r-390-bounces at mailman.qth.net [mailto:r-390-bounces at mailman.qth.net]
On Behalf Of AI2Q
Sent: Friday, June 10, 2005 1:19 PM
To: Peter Worrall, G4GJL; bernie nicholson; r-390 at mailman.qth.net
Subject: Re: [R-390] re cost reduction ect ect

No, sorry Pete, I have no recordings.

However, suffice it to say that the ability to listen to SSB stations, where
are some are weak and some are very strong, without having to twiddle any
knobs, really makes the outboard detector shine.

Vy 73, AI2Q, Alex in Maine
http://users.adelphia.net/~alexmm/ai2q.htm

----- Original Message -----
From: Peter Worrall, G4GJL
To: AI2Q ; bernie nicholson ; r-390 at mailman.qth.net
Sent: Friday, June 10, 2005 3:47 PM
Subject: Re: [R-390] re cost reduction ect ect

I wonder if you have recordings of SSB resolved barefoot with the 390 and
then with the adaptor added?
Many of us would be interested I suspect.

Pete
G4GJL
----- Original Message -----
From: "AI2Q" <ai2q at adelphia.net>
To: "bernie nicholson" <vk2abn at bigpond.net.au>; <r-390 at mailman.qth.net>
Sent: Thursday, June 09, 2005 5:25 AM
Subject: Re: [R-390] re cost reduction ect ect

> Hi Bernie:
>
> For your reference, you can see my R-390A product detector circuit at this
> URL:
>
> http://users.adelphia.net/~alexmm/Prod_det/detector.htm
>
> Vy 73, AI2Q, Alex in Maine
> http://users.adelphia.net/~alexmm/ai2q.htm
>
>