Date   
Re: ND6T AGC Troubleshooting

WA6PXC Rich <sigma@...>
 

Hi Jerry,
Thanks for your input. I agree, his circuit is elegant & simple .
I wanted to build an AGC and Squelch circuit on a separate board and just
remove one component & connect two wires to the BitX board (minimum mod to BitX board).
See attachment..
Rich

-----Original Message-----
From: "Jerry Gaffke via Groups.Io"
Sent: Jan 12, 2019 9:18 AM
To: BITX20@groups.io
Subject: Re: [BITX20] ND6T AGC Troubleshooting

Rich,

You are correct, it is better to have AGC reduce gain at RF or in one of the IF amps.
The audio pre-amp at Q70 of the uBitx (or Q16 of the Bitx40) is the weak link, 
restricting available dynamic range if AGC attenuation occurs after that audio pre-amp.

However, take another look at the ND6T AGC, it attenuates the RF as it comes in from the antenna.
Not unlike your design, but using FET's instead of op-amps.
Simple and effective.

Would be interesting to do an A/B comparison of your implementation vs the ND6T AGC.
Lots of variables here regarding dynamic range of the attenuation, attack and release times, distortion, ...

Jerry, KE7ER


On Sat, Jan 12, 2019 at 08:46 AM, WA6PXC Rich wrote:
Hi Armin,
I developed this AGC because all of the others act on the audio signal.
When the signal is really large, the audio is already distorted so we
just control a distorted signal. Please see the Attachment.
 
To get an idea of how well this type of AGC will work on your radio,
remove R-12 and replace it with the FET as shown in the schematic,
with just the 10K resistor going to the gate.
Using a 10K pot, connect it across 5v and connect the wiper to the Gate resistor.
As you turn the pot, it will vary the RF gain.
 
Have fun...
Rich

Re: Tracing receive signal for low signal

jim
 

K...Generally an RF generator output is "calibrated" with a termination connected ...Generator output 100 mv P/P open circuit ..when terminated into 50 ohm ..output "split"  half absorbed by generator, half by termination thus actual reading 50 mV ...

with 50 mV open circuit through 50 ohm generator + 1200 through test probe = 1250 ohm  ...then terminated into 50 ohm

voltage drop across 1250 ohm in series with 50 ohm termination ...roughly 25:1 means 1/25 of the signal is dropped across load, 24/25 dropped across generator/test lead
 
1/25 of 50 mV is about 2 mV  ...

You would do better with a random piece of coax and a blocking capicitor from the signal generator ...No coax?  use a twisted wire pair ...maybe noisier, but lots less loss ..short (2 - 3 feet of wire/coax)

Jim

On Saturday, January 12, 2019, 8:17:13 AM PST, Dave Space <davesspacebar@...> wrote:


Thanks for all the info.  Ok trying to see if I got this. It does look like loading.

1x probe is 1.2k
10x probe is 8.9 Mohms

I checked the probe directly outputting 50mv verified 50mv through a .01uF blocking cap @ 11.996MHZ
I put the generator into tp17 and measured at tp17 and I would say less than 5mv p-p was being read from what I could tell.

I wasn't understanding if there is a way to now match this better or not?

Thanks

Re: Tracing receive signal for low signal

Dave Space
 

Ok lets see if I did this right.

Setup:
Signal generator probe at 1x through 0.01uF cap at tp17
ch2 set probe at 10x at cap/tp17
ch1 set probe at 10x at tp20
Scope set to 1x for both channels
Signal generator set to 1v at 11.996MHZ

Measurements:
Ch2 measured 6mv p-p * 10 = 60mv being generated
Ch1 measured 5mv p-p * 10 = 50mv out of filter at 2khz frequency measured for approximately ~-1db loss

Does that seem about right?

(Its hard to adjust my frequency generator for best signal once set once set it can only change easily in 10k hz increments with the knob - I tried adjust a bit manually entering and that seemed to be where I get the best tone)




Re: Tracing receive signal for low signal

Dave Space
 

On Sat, Jan 12, 2019 at 10:48 AM, jim wrote:
 
K...Generally an RF generator output is "calibrated" with a termination connected ...Generator output 100 mv P/P open circuit ..when terminated into 50 ohm ..output "split"  half absorbed by generator, half by termination thus actual reading 50 mV ...
 
with 50 mV open circuit through 50 ohm generator + 1200 through test probe = 1250 ohm  ...then terminated into 50 ohm
 
voltage drop across 1250 ohm in series with 50 ohm termination ...roughly 25:1 means 1/25 of the signal is dropped across load, 24/25 dropped across generator/test lead
 
1/25 of 50 mV is about 2 mV  ...
 
You would do better with a random piece of coax and a blocking capicitor from the signal generator ...No coax?  use a twisted wire pair ...maybe noisier, but lots less loss ..short (2 - 3 feet of wire/coax)
 
Jim
 
On Saturday, January 12, 2019, 8:17:13 AM PST, Dave Space <davesspacebar@...> wrote:
 
 
Thanks for all the info.  Ok trying to see if I got this. It does look like loading.

1x probe is 1.2k
10x probe is 8.9 Mohms

I checked the probe directly outputting 50mv verified 50mv through a .01uF blocking cap @ 11.996MHZ
I put the generator into tp17 and measured at tp17 and I would say less than 5mv p-p was being read from what I could tell.

I wasn't understanding if there is a way to now match this better or not?

Thanks
Ah ok, sorry missed this message before I tested.  Ok let me see if have something else that will work with less loss... I might have a coax I can put into some alligator clips or wire leads or something.

Re: Tracing receive signal for low signal

Dave Space
 

Ok slightly modified setup with coax

Setup:
Signal generator into coax then some alligator clips 0.01uF cap at tp17
ch2 set probe at 10x at cap/tp17
ch1 set probe at 10x at tp20
Scope set to 1x for both channels
Signal generator set to 400mv at 11.996MHZ  (voltage input less here since clearly coax caused less loading and more output voltage)

Measurements:
CH2 measured 9mv p-p * 10 = 90mv
CH1 measured 7mv p-p * 10 = 70mv - saw ~2khz output tone
Estimated ~-2db loss


uBitX v4.3 calibration

Vic WA4THR
 

I see someone else asked this question before, but the thread wandered off into BFO settings, shipping issues, and other things and never resolved this that I can tell. I built a BitX40 last winter and was thrilled with it. I made some hardware changes and some software changes to the Amunter sketch and have had a blast with that very capable radio on SSB, both fixed and mobile. Now I am finally getting some time to start work on a uBitX and I have one that was delivered to me last summer and is a v.4 board. It powers up showing it is v4.3 and sort of worked (only trying receive at this point), but the audio sounded terrible and was fairly weak. I made a slight adjustment to the BFO that helped a little with the pitch, then noted that everything was 1kHz off frequency as compared to another rig. The posted alignment instructions on the HFSignals website seem to refer to an older version of the software. This is what another user observed:

V3:
Menu: "Setup on?"
Menu: "Set Calibration?" 
If button-
Print "Set to Zero-beat,"
Print "press PTT to save"

V4:
Menu:  "Settings on?"
Menu: "Set Calibration"
If Button -
Print: "Press PTT & tune"
Print: "to exactly 10 MHz"
delay -
Then it goes into TX - and the PTT Delivers a sidetone -

and this is similar to what I am seeing. If I tune to 10MHz, and I supposed to zero beat with another receiver tuned to 10mHz? If I ground the orange line labeled as PTT, nothing happens. So...??

=Vic=

Re: Unresponsive

Lawrence Macionski <am_fm_radio@...>
 

And I thought you just didn't loan money....

Re: Tracing receive signal for low signal

jim
 

Sounds close enough to me

Jim

On Saturday, January 12, 2019, 11:10:26 AM PST, Dave Space <davesspacebar@...> wrote:


Ok lets see if I did this right.

Setup:
Signal generator probe at 1x through 0.01uF cap at tp17
ch2 set probe at 10x at cap/tp17
ch1 set probe at 10x at tp20
Scope set to 1x for both channels
Signal generator set to 1v at 11.996MHZ

Measurements:
Ch2 measured 6mv p-p * 10 = 60mv being generated
Ch1 measured 5mv p-p * 10 = 50mv out of filter at 2khz frequency measured for approximately ~-1db loss

Does that seem about right?

(Its hard to adjust my frequency generator for best signal once set once set it can only change easily in 10k hz increments with the knob - I tried adjust a bit manually entering and that seemed to be where I get the best tone)




Re: Unresponsive

Dexter N Muir
 

No, we don't. What we do is communicate: giving (and recieving) advice on-topic. The old 'saw': give a man a fish ....
Hopefully helpfully,
73 de ZL2DEX

uBitx ver 1 for sale

Hank W6IR <w6ir@...>
 

I have too many other projects and 2 very active grandkids so something has to give. So I'm selling my Version 1 board (#692) along with the following items.
Case with connectors installed. Bezel for the 2 line display. W0EB serial lcd board (no serial display just the board with Nano installed).
External connector for the nano on back panel. Volume control pot has on/off switch.
It is wired up and working.

Looking for $140 shipped priority mail to USA.
Contact me at my QRZ address.

Hank
W6IR

Re: Mini PA für my uBITX #ubitx

AA9GG
 

LOVE your BitX "twins"!! 

Do you have a link for the PA module on ebay?  I see tons of the MRF186 but no pcbs for it...



Virus-free. www.avast.com

Re: Tracing receive signal for low signal

Dave Space
 

Just to see if using coax would have changed other numbers on the signal generator
I retested (I might have measured wrong before) input took 700uV in at the antenna got to 100mv peak to peak at vol high same as recently with my pre-amp installed. So no difference with coax at that location.

As mentioned before while it seemed there was crazy loading when I used higher voltage like 700mv it had more difference than Evans board (though he input at only 500mv). 
Inputting at lower voltage at the antenna at 30mv @ 7.147MHZ I got about 14-16mv out at TP1.  Removing my jumper and injecting at the K1 side it was just slightly more at L4 maybe 15-18mv.  So the K3 relay doesn't seem to have a huge effect.  

So given there doesn't seem to be any significant loading there.  So don't think it makes sense to cut the trace to the mixer anymore right to test?

Evans board seems to match a lot of the voltage levels on parts so seems like something else is up perhaps.
It seems like we do know with enough power things will work... so potentially could just add amplifier stages after Q12 and after Q32 would likely workaround the rather anemic receive?  

(Diagnostics definitely seem harder when things sort of work just not quite as well as you'd like... especially when the measurements are in microvolts that most instruments won't be able to measure well with all the noise in the system/on the board... and using larger voltages for some reason is inducing more loading as per the earlier measurements.)

Re: uBitX v4.3 calibration

Curt
 

Vic

Something very similar happened to me trying to do frequency calibration. Yes 7 got it into transmit also. Somehow I did improve it.

Very important,  it was slow to dawn on me the display reflects ssb operation. When we go into cw, and see ourself on RBN note the frequency will be about 700 hertz higher or lower, depending upon whether we are in usb or lsb. The kd8cec website has a nice explanation here.

My board came with the bfo way off. With the nice narrow xtal filter, the bfo must he close to get much ssb energy into the passband. Patiently work this before thinking something is amiss. I needed to visit the web to find a typical bfo value to get close enough to set it. The ubitx receives wonderful with bfo aligned.

Curt

Mike Short
 

Powered up bitx40 and immediately have a tone in the speaker, and no info on display. 
I initially miswired the Raduino. I used the 8 pin connector and matched the colors, didn’t use the 16 pin connector.

When it powered up the display showed briefly. I powered it off. I rewired it and have the tone. 

mike

JackAl Voltage Chart

Jack, W8TEE
 

All:

The voltage chart for the JackAl board is now posted on hamradiodesigns.com. The program source code will be available for download on Monday.

Jack, W8TEE
Al, AC8GY


Re: uBitX v4.3 calibration

_Dave_ K0MBT
 

Yeah Vic you want to calibrate the frequency display first. I didn't like using wwv for calibration of frequency. They have the intermittent carrier. I reloaded to a version 1.08 of kd8cec and tuned it to a Commercial AM station I figured a big station would have a much more accurate carrier than I would need. Once you do that then go back in and set the bfo to 11996500 or so and adjust it from there.

I tried the latest version of kd8cec, 1.1, but didn't find the option to run it on the 1602 display,

Vic see you down the log.
73
Dave
k0mbt

Re: Tracing receive signal for low signal

jim
 

Random cutting of traces is not "my" way of trouble finding ..tho at times I've  been tempted ..Amps at Q12 and Q32 would both need to be well terminated as they would be terminated into the crystal filters ..Both (any) crystal filter is highly reactive around the passband frequency and as such need to be well matched/terminated ..Easiest to add amp at antenna end 50 ohm in/out ..use relay or switch to control receive/transmit ..

That being said, I've dead-bugged a version of the Q30/31/32 amp ..replacing Q31 with a PNP (2n3906).. R31=10 k and R34 150 ohms.. collector of Q31 to R35 ,,,gives (at first glance) ~50 db gain ... Have yet to hack the microBitx to really try it tho

Expect it will be noisy and unstable , but ...maybe

Jim


On Saturday, January 12, 2019, 5:55:15 PM PST, Dave Space <davesspacebar@...> wrote:



So given there doesn't seem to be any significant loading there.  So don't think it makes sense to cut the trace to the mixer anymore right to test?

Evans board seems to match a lot of the voltage levels on parts so seems like something else is up perhaps.
It seems like we do know with enough power things will work... so potentially could just add amplifier stages after Q12 and after Q32 would likely workaround the rather anemic receive?  



Groups.io Links:

You receive all messages sent to this group.

View/Reply Online (#63842) | Reply To Group | Reply To Sender | Mute This Topic | New Topic


Re: Tracing receive signal for low signal

Dave Space
 

I had that pre-amp earlier in the receive line between K3 and K1 and what I noticed that amp vs the new pre-amp I put in at volume high  is the pre-amp between K3 and K1 just sounded like it amplified the noise as much as the signal making the noise as loud as the signal.  Where the amp at vol high seems to have done a better job picking up the signal and less noise (at least if I don't have the gain too high).
I think maybe because the low pass filter rejected a bunch of noise so the amplification stages amplified less of the noise (at least that's the best way I could describe it from a noise quality I can hear).  Thus why I was thinking it might be better to amplify after the low pass filters (it probably lost some signal there also anyway).

What I'm thinking is two transistor stages common emitter then emitter follower to get the low output impedance for following Q12 (so common emitter + emitter follower )and Q32 (common emitter + emitter follower)

The only other way I can think of to track down the board differences is buy a second uBitx board that hopefully works (though I'd be hosed if it didn't) and just analyze component by component all the way through until I find something with a meaningful difference.

Re: Tracing receive signal for low signal

Dave Space
 

Actually I think with the pre-amp I might have had a loose connection and had to adjust the gain.
It seems a lot better now.  I actually got around 130mv p-p at vol high with 100uV into the antenna.  I might have been getting inconsistent readings due to the breadboard loose connections or something.

Also often I would do some initial on air tests with a ham stick before switching over to my full 40m dipole.

I need to do more testing but the on air tests sound really good with this pre-amp in place now and connections good.

I might need to actually dead bug the pre-amp onto an actual board to keep the connections from being loose.

(at the moment I'm listening to a s7-s9 signal over the earbuds and I'm not wearing them they are on the other side of the room :) )

Re: Tracing receive signal for low signalou've

jim
 

If you've got it working to your satisfaction, thats probably a good place to stop ..

Jim


On Saturday, January 12, 2019, 9:43:00 PM PST, Dave Space <davesspacebar@...> wrote:


Actually I think with the pre-amp I might have had a loose connection and had to adjust the gain.
It seems a lot better now.  I actually got around 130mv p-p at vol high with 100uV into the antenna.  I might have been getting inconsistent readings due to the breadboard loose connections or something.

Also often I would do some initial on air tests with a ham stick before switching over to my full 40m dipole.

I need to do more testing but the on air tests sound really good with this pre-amp in place now and connections good.

I might need to actually dead bug the pre-amp onto an actual board to keep the connections from being loose.

(at the moment I'm listening to a s7-s9 signal over the earbuds and I'm not wearing them they are on the other side of the room :) )