Retro 75 Modulation Frequency Response

[k1swl]

Guys-

I'll introduce another 'wild card' into the ongoing discussion.    I just received the following message via e-mail:

>>Completed the Retro 75.  Traced a microphone/modulation problem to R28.  The bottom of R28 is floating and not grounded.  I inspected the board under a stereo microscope and appears that the pad goes nowhere.  I grounded R28 and now all is well.
<<

Uh.... he's right! R28 is flapping in the breeze.  It certainly would affect the ALC dynamics, but I haven't had a chance to look at it.  I did ask the gentleman for further information on the before-and-after performance.


with egg-on-face-  73, Dave- K1SWL

[kg8lb]

 Dave,

  Thanks for the heads up. I just checked the board on my yet un-built Retro #3 and R28 is indeed floating at one end, the other end of R28 going to Q 10 as it should. I don't have a stereo microscpoe but a back lit x-ray type view of the board agrees with an open ohm meter reading.
   I will check the other two (Both seem to be working quite well) tonight .  
No problem, this stuff happens and the little bugger seems fine. In any event the fix is quite simple.

[k1swl]

I don't know what effect grounding R28 properly will have. Intuitively, it'll raise the dynamic threshold and thus increase the modulation percentages.

There was a resistor (R27) value change a while back on the SWLRetro-75 user group, now incorporated in the present configuration.  It had the effect of slowing the attack time, and probably is interactive with this R28 discovery.

73- Dave, K1SWL

[k1swl]

In response to my question about R28, here's the response:

>>I checked the rf waveform on my scope and wasn't able to get full modulation.  At first I thought it was the Radio Shack microphone, so I went to an amplified mic, but increasing gain of the mic caused distortion, but no increase in modulation.  I was seeing a good waveform at the output of U5, but little audio at the junction of C27 and C28, so I pulled Q9 to no effect and then Q10 which did have an effect.  This led be to R28.  I suppose the magnitude of the problem would depend upon the leakage of C32.  I may just have been lucky enough to get a low leakage capacitor.
I now have full modulation without distortion.
<<

(Thanks, and a tip of the hat to Robert Sherwood for his followup)  73- Dave, K1SWL

[w3jn]

All I can say, Dave, is thanks much for offering this kit.  It's certainly stimulated a lot of interest. 

It can't be easy designing the kit, buying the parts, and you're certainly not getting rich off the profit margin.  Kudos for a job well done!

[kg8lb]

All I can say, Dave, is thanks much for offering this kit.  It's certainly stimulated a lot of interest.  

It can't be easy designing the kit, buying the parts, and you're certainly not getting rich off the profit margin.  Kudos for a job well done!

Times two !
 It is amazing how many contacts I make with people who say "I have one of those kits. I just haven't put it toghether yet." Made a lot of Rtro to Retro contacts as well.
   Thanks to Dave and Robert for the updates.
And... Thanks to the AM Forum for providing a place to exchange information .

[KD6VXI]

Although I haven't (yet) bought and built one, if this is your biggest error, Dave, you're doing GREAT!

--Shane
KD6VXI

[WD5JKO]

Just waking up here getting ready for my shift..

  I know in my effort with my retro that the AGC is working, and at one point I clipped onto R28 (Q10 side) to add an R-C to ground to increase "decay time" without slowing down the attack time. It was clearly working, but I took it out. Maybe if the other side of R28 is floating, then the AGC behavior will be less abrupt once it is corrected.

  Thanks for the input, I will get to this in the coming weekend.

   For those who want to Hot-Rod their retro75, this thread still applies. The R28 issue once corrected (jumper) does sound like it validates the stock design. As for me I liked the V8 conversion to the Chevy Vega.. 

Jim
WD5JKO

[k1swl]

Jim-

Thanks for all the good work on the audio response!

You asked a while back about the number of Retro-75s out there.   By today's count, I've shipped 375 of them so far.  Not bad for an 'obselete' mode!

When I first mentioned my plan to do an AM rig to a good friend, his reaction was "But there's only 4 guys on AM, and they're all running broadcast stations!"  'Au contraire!'  I'd have been satisfied if I'd used up the initial lot of 100 boards, but that didn't take long.

I now have a spectrum analyzer on order- I've been too long without one.  I hope to wrap up the 'Retro-Helper' in the next several weeks and need the S-A to verify some recent changes for spectral purity.   The little doodad mounts on the rear wall of the enclosure (that's what the spare holes are for). The transmitter will track the receiver tuning-  full VFO control.   

73- Dave

[kg8lb]

 Oh boy !  The retro 75 is a winner for sure. AM ops are naturals for building their own gear , kits or homebrew. The Retro Helper should be a huge seller as well. Is there still a chance for the 40 meter Retro 40 ?

[k1swl]

The Retro-40 is at the top of my round-tuit list. I need to modify a spare Retro-75 and listen on 40M for a while.  I'm fairly sure the Retro will not stand up to 'hot' band conditions on 40M, but it may be fine for daytime use.  (It's a matter of IMD from SWBC 'powerhouses').

I did get started on one- the Tx put out 1.85W carrier, so that looks OK.

It looks like the IF would be 5.185 MHz-  I wouldn't even need to change the LO components or L2's turns count.

73- Dave, K1SWL

[K1JJ]

Dave,

Glad to see you've joined the forum.

I've been impressed with your product and still scratch my head when considering the cost of parts and how little you charge.  Your efforts are certainly appreciated here.

Thanks, OM.

Tom, K1JJ

[W1RKW]

Tom,
You'd be amazed at Dave's tech support too!

Dave defines the ole tried and true ham spirit.

[k1swl]

Tom-

Good to hear from you!  We lived in the NW corner of Colchester a while back. I'd breadboarded the receiver portion of the Retro before we moved, and needless to say, I could copy you really well. Wish I'd been further along with the project, or I could have easily worked both you and Bob, 'RKW.

73-  Dave

[w3jn]

Dave - in terms of IMD response for your 40 meter version, a few possibilities come to mind -

1 - get rid of diodes D1-D4.  They'll definitely act as a mixer and create all kinds of undesireable crap.
2 - consider one more stage of bandpass filtering at the input if you get images
3 - a 3 dB or 6 dB attenuator can work wonders with probably minimal effect on S/N.  Perhaps this would be an option on the circuit board, or on a little daughterboard that a guy could add if he gets IMD.  The SA-602 gives pretty amazing performance for what it is, but a gigantor antenner could certainly overload it, considering the close proximity of blowtorch signals on 49 and 40 meter broadcash freqs.

[k1swl]

Johnny-

Thanks for the thoughts!

D1-D4, along with L1 and C1, are doing the T-R switching at the front of the receiver.  There's 4 of them there to improve on the performance over a single pair of the diodes. Elecraft uses a slightly different scheme- a transistor conducting to ground on the positive half-cycle and using B-E breakdown on the negative half-cycle.  I don't remember the results I got with intercept point calculations (it was too long ago)- but I suspect the '612 wasn't far behind it as the next weak link in the chain.  For that matter, if removing the diodes makes a substantial difference, a small T-R relay would do the job. 

Yes- a few dB of attenuation up front might work wonders, since the intermod is a higher-order effect.   I'll know once I have a chance to give it a listen.    I'm not sure yet about any image issues.  It wasn't a concern with the 75M version, with the image at about 8.1 Mhz and therefore well down the attenuation curve on the output low pass filter.   

73- Dave, K1SWL

[N4LTA]

Dave,

Thanks again for this kit and all of the others you have done.  As you and I both know, selling kits can be very challenging and the challenge is mostly the "non technical"  part.

It was nice talking to you again when I bought the two kits several months ago. Keep up the great work.

Pat
N4LTA

[Ralph W3GL]

  Dave,

  While I'm not a customer of any  of your products, I have looked at 
  the projects you have provided in the past...
 
  Seams to me a Retro-160 would be a better item than the 40 meter
  version.  Just my opinion however both the 40 or possibly 160 should
  reach a good audience.

  Logic says 160 faces less potent flamers than the 40 meter version.

[k1swl]

Ralph-

Thanks for your comments.  The last time I was on 160M AM, I was in high school...and it was still the vacuum tube era.  From what I remember of operation on 160M, activity was pretty well distributed across the band.  The challenge with something as simple as the Retro would be to cover a big swath of the band without repeaking the tuned circuits.   

If AM activity's now pretty well centered around one particular frequency, that'd be a different matter.  :-)    73- Dave, K1SWL

[kg8lb]

Forty meters is a good band for fairly long distances on low power. Forty meter antennas are easily carried for portable use and a 40 meter RX on AM is handy for plain SWL use when ham activity is not an option.

[KX5JT]

I vote for 40!

[Steve - K4HX]

Since most AM activity on 40 meters is during the daytime, broadcast stations are not much of an issue. The propagation and the antenna requirements favor low power operation over 80 or 160 meters. So, 40 would be my vote.

[KX5JT]

Since most AM activity on 40 meters is during the daytime, broadcast stations are not much of an issue. The propagation and the antenna requirements favor low power operation over 80 or 160 meters. So, 40 would be my vote.

It's too bad because 40 is a great band early mornings with plenty of space between broadcast stations, but rarely any AM to be drummed up.

[WD5JKO]

Update,

What I noticed though from the beginning when I first put a tone into my Retro-75 when 100% original, was that I was limited to ~ 30% modulation with a sine wave, and the waveform was not symmetrical once limiting kicked in. With speech I could kick up to maybe 60-70% on peaks.

   So last night I grounded the floating end of R28. In my case the AGC circuit is different than stock (see earlier post), and the result was not very apparent. Maybe that was because I already did have a voltage divider in my arrangement. That resistor R28 should discharge the peak hold capacitor (C32) at an RC rate, and without it the cap would stay charged to a voltage just shy of base current on Q10, and get a boost on audio peaks turning on Q10. In stock form, hooking up R28 as intended should delay the onset of AGC, and that is a good thing.

   So whether were stock, or as I have modified, the AGC system seemed prone to cause asymmetry of an incoming symmetrical wave, and also could be easily overdriven such that hard limiting could occur, and only occur on one polarity of the incoming signal. I mentioned this in my last update. Looking at Q10, and how it was used made me think about a soft clip instead of a hard clip. So with one of those 4 am inspirations, I clipped the collector of Q10 at board level, and soldered a resistor from the collector pin hanging to the trace the collector was formally hooked to. I tried a few values, and settled on 1.5K. I don't fully understand what happened, but all the bad AGC behavior I was mentioning just went away! I can run a sine wave fully symmetrical well into limiting now, and with speech I can see peaks to 150% before the TDA2003 modulator clips.

   I hooked the scope in X:Y mode configured for Trapezoid mode, and that RF power FET modulates with a perfect trapezoid up to 150%. So with 4W carrier (14V B+), 150% modulation, I'm up to 25w PEP now.

   In the circuit I posted (Rev - A) there are a few changes:

1.) C33 was installed as per silkscreen, but in my rewiring changes, that was backwards (+), and I changed it to 330 uf
2.) The R-C (270-10uf) across the modulated B+ was changed to 150 ohms - 47 uf. This is the passive keep alive supply that works well without needing a bias supply or additional diodes.
3.) I moved the adjustment pot for AGC threshold to the junction of R27 and D10.

  At this point, having control of the asymmetry of my voice (> 1.5:1) is essential. I have two Dynamic microphones, and both were phased opposite to what I need. Fortunately it was simple to swap the wires on the mic cartridge to fix this issue. I was thinking of how we can do this within the Retro-75. Well it appears that half of U5 (dual op-amp) is unused, so adding a simple X(-1) inverting amplifier could be added with a SPDT switch to select speech polarity could be used.

  I also have a condenser mic from the XYL's tape recorder. So I tried that, and added a 4.7K pull up resistor from 6V to the mic input. The mic worked fine, but the polarity was wrong again.  
Here I need the inverting amplifier since a condenser mic polarity cannot be changed.

  So does the asymmetric nature of the male voice vary a lot person to person as well as the dominate phase? With three microphones I am 0 - 3!  

EDIT: Found my Shure Unidyne II dynamic. it too is phased backwards! So am I like the left handed guy in a right handed world? Now I'm 0-4 on mic polarity...

  I hope to make a PDF of these changes, and provide a link to my web space of these changes. That will take a while to do though.

  The modulation peaking 150% was reported as somewhat fuzzy by a few stations, while others said it was loud and clear. This is to be expected. I should probably limit to 120% by turning the pot. Still I held my own in a big Texas round table this morning with stations in west Texas, Houston, and Dallas all hearing me fine from the Austin area.
  
Jim
WD5JKO  

[w5omr]

Still I held my own in a big Texas round table this morning with stations in west Texas, Houston, and Dallas all hearing me fine from the Austin area.

Since I dwell on a cliff now, and no way to monitor the bands, I wouldn't feel it a bad thing were you to call me on my shoe-phone (invoke mental image of Maxwell Smart) and say "hey, get on the air!" and I'd go get in the truck and make a mobile contact.