Mod for more AF gain on K2

[Mike S-8]

It seems that some people have problems with low audio out on some K2's. There was a recent thread where I mentioned that it may be possible to easily substitute an LM386 for the LM380 AF amp and get more gain. I'm happy to report that this works FB.

In actual practice, I used an NJM386BD, available from Mouser for about 30 cents.

The 386 has a slightly different pinout than the 380, and also requires an external component to set the gain (the 380 has a fixed gain of 50). Pins 1 through 4 are identical, but pins 5-8 differ slightly:

Pin..380......386
5....Gnd......Vout
6....Vout.....Vss
7....Vss......Bypass
8....Bypass...Gain

As you can see, three of these pins are simply shifted by one position. I installed a socket where the 380 was (an AMP one, I've never had problems with either these or the nice Augat machined ones). I was then able to slightly bend the pins on both sides of the 386 slightly to offset them and allow all of the common pins to be plugged into the socket. This leaves the Gain pin (8), which simply needs to be connected to pin 1 via a 10 uF cap (sets the gain to 200), which was done piggyback style. Even with the socket and cap, there was still plenty of clearance for the KAF2 (and presumably for a KDSP2).

Results? At an AF gain setting which with the 380 produced 1.15 VAC at the internal speaker, I now get 4.00 VAC, an additional gain in power of about 11 db. It gets almost painfully loud with an S9 signal, and plenty loud with a signal which doesn't even register on the S meter. (I didn't have a problem with lack of power before, but now my AF gain goes to "11" HI HI).

If you compare the specs, the 380 can put out more power. In reality, I measured that the 386 begins clipping at about 4 V out and the 380 at about 5 volts, so the 380 only gets about 2 db louder. Having more output power available does no good if there's not enough gain to get you there. The 380 can go louder in absolute terms, but the 386 can get _very_ loud with _much_ smaller signals.

Mike

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[wayne burdick]

On Oct 22, 2004, at 1:17 PM, Mike S wrote:

> It seems that some people have problems with low audio out on some
> K2's. There was a recent thread where I mentioned that it may be
> possible to easily substitute an LM386 for the LM380 AF amp and get
> more gain. I'm happy to report that this works FB.

Mike,

I'm glad that this worked for you.

However, because of the problem with clipping at high signal levels, I
definitely don't recommend that anyone do this unless they're desperate
for more audio. The LM386 is a lower-power part and is also noisier
than the LM380, partly because of all that extra gain.

AF output can be increased in less drastic ways. Here are my favorites:

- On older K2s, make the 2nd XFIL modification. This allows you to
properly align L34, minimizing loss through the 2nd crystal filter in
both SSB and CW modes.

- Verify that the band-pass filters and L34 are peaked correctly.

- Use Spectrogram to properly align the BFOs as specified in the manual.

- Adjust the AGC threshold so that it isn't activated by low-level band
noise.
   (First, turn AGC OFF by holding PRE and ATTN. If the background noise
comes up,
   you may have the AGC pot set too aggressively.)

- Add the KDSP2 option, which provides programmable AF gain well above
the standard level.

- Add the KAF2 option, which also increases AF gain (and it can also be
adjusted by changing a couple of resistors).

- The gain of the preamp can be increased by a few dB with little
overall change in RX performance. Simply change R76 from 10 ohms to
about 5 ohms. You can substitute a 5.6-ohm resistor, or just parallel
another 10 ohm resistor across the existing one. In fact you can do
this "live" to see what the effect is.

- If all else fails, signal trace through the receive chain as
described in appendix E. You could be losing RX signal in the T-R
switch or elsewhere.

Note that you can check your K2's performance easily with an Elecraft
XG1 1-microvolt receiver test oscillator.

73,
Wayne
N6KR


---

http://www.elecraft.com

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[G3VVT]

Another alternative to fit the 386 in the place of an LM380 audio amp in  the
K2 would be to cut the print on the affected pins and fit insulated wire  
straps to change the print layout. Have not checked the 386 pin out to confirm  
the details, but pin 5 would need isolating from ground on the top/component  
side of the control board and on the underside 4 lines of print to pins 6, 7, 8
 would need to be cut with something like a Dremel tool. New insulated wire  
straps could then be added to change the print to that required for the  386.
 
My own K2 (serial no. 4168) does not suffer from low audio and is run with  
the AF gain control at 1/4 to 1/3 rotation from minimum for most uses  with
either the normal internal speaker or a single earpiece headset/electret  mic
that I use with the K2. Is mostly used for SSB with OP1 filter in use,  only
resort to using the "drum" (CW) in extreme circumstances to complete a  contact
when condition are poor. I would be more susceptible than many to the  effects
of low audio after loosing most of my hearing on one side after an  accident at
work and the remaining ear down to telephone bandwidth only.
 
Is the low audio complaint applicable only to early K2's, the effects of  
narrow filters in CW reception or is it a random effect caused by gain  variation
of active components?
 
Regards,
Bob, G3VVT
_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Mike S-8]

At 06:42 PM 10/22/2004, wayne burdick wrote...

>However, because of the problem with clipping at high signal levels, I definitely don't recommend that anyone do this unless they're desperate for more audio. The LM386 is a lower-power part and is also noisier than the LM380, partly because of all that extra gain.

I was able to make both clip, but in both cases it was when the volume exceeded a comfortable level. The LM386 was putting out over 1/2 W when it did.

I don't see where noise is characterized on the National datasheet, and nothing to indicate that the 380 has a better signal to noise ratio, which is what's really relevant. If more audio gain is needed to get useful volume, amplifying noise is unavoidable. The LM386 gain is adjustable, so if less than a 12 db boost is needed, that too is possible. It's reasonable to assume that a 386 configured for the same gain as a 380 would have comparable noise.

I agree with your point that adding gain (or decreasing loss) earlier in the chain is preferred, and offered this as a solution for when that fails to achieve adequate volume, as it apparently has for some users.

"Low power" is relative. I measured 4 VAC (p-p, ~500 mW ) on the 386 and 5 VAC (~850 mW) on the 380 at the onset of clipping, a relatively minor difference and inconsequential given that either level provides more than ample volume with the internal speaker. Going past clipping adds volume at the expense of distortion, but at the levels I observed it would only be useful for listening from across the room (or from the next room), where that may be a reasonable tradeoff. I doubt anyone who can make either clip is complaining about low volume.

Mike

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[wayne burdick]

I think we're on the same page, Mike. I just want to make sure builders
know there are other alternatives before they jump into a different AF
amp IC. Lord knows I've designed LM386's into a lot of rigs (the NorCal
40A, Sierra, SST, K1, and KX1, to name just those which made it out of
the lab). I tame the noise by putting a resistor/capacitor network in
series between pins 5 and 8. But ideally the gain and NF ahead of the
amp would be adequate so that no heroics are required here.

Regarding "how loud is loud," the place where you need the audio levels
that the '380 is capable of is in mobile or high-noise outdoor
situations--such as Field Day with a generator nearby. Or in my rickety
VW van at 60 MPH  ;)

73,
Wayne


On Oct 22, 2004, at 5:06 PM, Mike S wrote:

> At 06:42 PM 10/22/2004, wayne burdick wrote...
>
>> However, because of the problem with clipping at high signal levels,
>> I definitely don't recommend that anyone do this unless they're
>> desperate for more audio. The LM386 is a lower-power part and is also
>> noisier than the LM380, partly because of all that extra gain.
>
> I was able to make both clip, but in both cases it was when the volume
> exceeded a comfortable level. The LM386 was putting out over 1/2 W
> when it did.

---

http://www.elecraft.com

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Mike S-8]

At 11:00 PM 10/22/2004, wayne burdick wrote...
>I tame the noise by putting a resistor/capacitor network in series between pins 5 and 8.

Perhaps a "real" LM386 has more noise. Noise in the NJM386B I used measured ~20 mV p-p in the absence of signal (AF pot at min, measured at the speaker). That's 46 db down from full output, just barely audible via either speaker or headphones. This noise is quickly concealed by other rig noise (preamp off, dummy load connected) as soon as you turn up the AF just a bit, except for some higher frequencies which don't get masked and are only noticeable with headphones. Putting the 380 back in, I measured 8 mV of noise, about 8 db better than the 386. That difference contrasts favorably to the 11 db difference in gain I observed between the two.

I then installed an RC network with the 386, as mentioned above (2.7K/.01u, copying what's in the KX1). It didn't change the level of the overall noise significantly, but it did pretty much remove the noticeable higher frequencies from the output.

All in all, not a bad change - trading about 2 db in maximum volume for about 11 db in gain. Since I socketed this and built everything piggyback on the 386, it's a simple chip swap to revert to stock condition.

_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com

[Leigh L. Klotz Jr WA5ZNU]

Although I had adjusted it to 3.80v, when I checked the AGC voltage, I
found it had crept up above 3.9v.
Since the adjustment happens early on in construction, it's likely that
the pot got bumped later.
This sequence of calibration could be a cause of reported low audio
level in 3000+ K2's.
I would recommend adding a reminder in either Part II or Part III
Alignment to check the AGC voltage.

Right now I have mine set at 3.50v but reading Don Wilhelm's message of
07/30/04, that may be a bit too low.
I can, though, confirm Don's assertion that listening for a change in
band noise with AGC Off is best done on a high band, not on 80 meters.  
(Ouch!)

73,
Leigh WA5ZNU

Wayne Burdick wrote:

> AF output can be increased in less drastic ways. Here are my favorites:
> ...
> - Adjust the AGC threshold so that it isn't activated by low-level
> band noise.
>   (First, turn AGC OFF by holding PRE and ATTN. If the background
> noise comes up,
>   you may have the AGC pot set too aggressively.)


_______________________________________________
Elecraft mailing list
Post to: [hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft   
Help: http://mailman.qth.net/subscribers.htm
Elecraft web page: http://www.elecraft.com