¿ªÔÆÌåÓý

ND6T AGC S Meter Calibration


Mark M
 

I have Don's & Kees' AGC board in my uBitX and am trying to get the S meter function in Ian's firmware working. I've been using the S meter helper in the Manager to try to set some meaningful values but have not been terribly successful so far. It seems like only the strongest signals give a usable level, it shows either a very small value or about a 75 value with little variation in between, it's either S1 or S9+ with nothing in between.

I borrowed an Elecraft XG2 receiver test oscillator from a friend to try the calibration with known input levels (it puts out 1uV and 50uV signals on 80, 40, and 20M). Measuring the voltage on the AGC board at J1 I get the following (I hope this is readable):

Band No Sig 1uV 50uV
80 .4mV 1mV 1.6V
40 .3mV .5mV 1.7V
20 .3mV .7mV 1.6V

Are these values typical? I'm not 100% sure the board is working correctly anyway...it's my first attempt at surface mount assembly so I may have messed up something altho I don't see any obvious problems. The radio seems to work just fine otherwise and the receiver seems to have decent sensitivity, to my ear at least.

The S meter is certainly not essential but since it's there, I'd like to see if I can get it working.

Anyway, any advice/opinions/comments are appreciated.

Mark AA7TA


 

Hi Mark.
I cant help, but I am in the same boat. I have built 3 (1 diy, 2 kit) of these boards in 2 different uBitx, and I don't think mine are working... On local nets I still get blasted by close by stations and struggle to hear distant ones, without twiddling the VOL...

I will watch this thread for replies.

Good luck, 73 Nick VK4PLN


 

The s-meter wire goes to the? pin at the hot end of C4? where the pins are.

Here is what I got, don't recall the frequency.??
.4 volts for 50 uV or -73 dbm
4 volts -40 dbm
3.9 volts -50 dbm
1.8 V -60
.55 V -70
.1 V -80

Using the following in the manager.
4,36,71,106,140,175,212,217

50 uV shows 3 bars on the meter
-63 dbm shows 6 bars
-57 dbm shows? +
-51? dbm shows ++

I was using an IFR 1200s and Fluke digital multimeter.? The s-meter pin was disconnected when doing the voltage measurments.

de ku4pt


On Mon, Jul 23, 2018 at 7:54 PM, Mark M <junquemaile@...> wrote:
I have Don's & Kees' AGC board in my uBitX and am trying to get the S meter function in Ian's firmware working. I've been using the S meter helper in the Manager to try to set some meaningful values but have not been terribly successful so far. It seems like only the strongest signals give a usable level, it shows either a very small value or about a 75 value with little variation in between, it's either S1 or S9+ with nothing in between.

I borrowed an Elecraft XG2 receiver test oscillator from a friend to try the calibration with known input levels (it puts out 1uV and 50uV signals on 80, 40, and 20M). Measuring the voltage on the AGC board at J1 I get the following (I hope this is readable):

Band? ?No Sig? ? 1uV? ? 50uV
?80? ? ? .4mV? ? ?1mV? ? 1.6V
?40? ? ? .3mV? ? ?.5mV? ?1.7V
?20? ? ? .3mV? ? ?.7mV? ?1.6V

Are these values typical? I'm not 100% sure the board is working correctly anyway...it's my first attempt at surface mount assembly so I may have messed up something altho I don't see any obvious problems. The radio seems to work just fine otherwise and the receiver seems to have decent sensitivity, to my ear at least.

The S meter is certainly not essential but since it's there, I'd like to see if I can get it working.

Anyway, any advice/opinions/comments are appreciated.

Mark? ? ?AA7TA





 

Mark,
If you do not require calibration to a standard I would suggest you set your lowest value (mine is set to 2) and your highest value (you measured S9 at 50 uV in that gives you 1.6 V) at one third of 254 (82 to 85) and select one of the samples. I liked sample 3 and my highest value is set to 80.
The software will calculate intermediate values for you.
Encode and write to uBitx.
73,
--
Ion

VA3NOI


Jack Purdum
 

Are we rearranging the deck furniture on the Titanic here? How important is it to provide an S meter reading down to the last dB of the log scale? I recently worked a contest running 80W and I was a 5-9 EVERYWHERE on the planet! If someone asks, I'll give an S meter reading that comes off my rig's S meter, but I have no clue how accurate it is. Is a 59 from me the same as the 57 report from someone else? I don't know. If I listen to signals on my BITX and compare its S meter reading to my commercial rig and they are the same at 59, I think that's good enough for most of us. If you're doing antenna or some other performance-related research for an article or something, I can see the need for a more precise measurement. But for day-to-day QSO's, a quick calibration against my commercial rig is probably good enough.

Jack, W8TEE

On Monday, July 23, 2018, 9:01:03 PM EDT, Ion Petroianu, VA3NOI <ion.petroianu@...> wrote:


Mark,
If you do not require calibration to a standard I would suggest you set your lowest value (mine is set to 2) and your highest value (you measured S9 at 50 uV in that gives you 1.6 V) at one third of 254 (82 to 85) and select one of the samples. I liked sample 3 and my highest value is set to 80.
The software will calculate intermediate values for you.
Encode and write to uBitx.
73,
--
Ion

VA3NOI


 

S-meter readings won't mean much until we all standardize our antennas.
Officially, an S-meter measures the signal power going into the radio.
But of course, all the cool radios have one so the uBitx wants one too.

An S-Meter using the diode detector of the AGC board will not have much range.
Given the two 1n4148 diode drops involved and 5v max into the Nano, I'd expect a range of under 3 S-Units.
At 6dB per S-unit, that's a voltage ratio of? ?10**(3*6/20) = 7.94.??
Use a log amp like the AD8307 if you really want an S-meter.?

Jerry, KE7ER


On Mon, Jul 23, 2018 at 06:10 PM, Jack Purdum wrote:
Are we rearranging the deck furniture on the Titanic here? How important is it to provide an S meter reading down to the last dB of the log scale? I recently worked a contest running 80W and I was a 5-9 EVERYWHERE on the planet! If someone asks, I'll give an S meter reading that comes off my rig's S meter, but I have no clue how accurate it is. Is a 59 from me the same as the 57 report from someone else? I don't know. If I listen to signals on my BITX and compare its S meter reading to my commercial rig and they are the same at 59, I think that's good enough for most of us. If you're doing antenna or some other performance-related research for an article or something, I can see the need for a more precise measurement. But for day-to-day QSO's, a quick calibration against my commercial rig is probably good enough.
?


Mark M
 

Guys...I understand that S meters are mostly subjective and it's certainly not a necessity but everything else works so well and the capability is there so why not try to get it to work? I'm not looking for mil-spec accuracy, I was just wondering if what I see is typical or if it indicates a problem somewhere in the AGC board. I don't think I've ever had a qrp rig with an S meter so it's kind of icing on the cake.

And like Jack, I've noticed that for some reason, during contests they all seem to indicate S9. ;)

Anyway, thanks for the inputs...

Mark AA7TA


Jack Purdum
 

Mark:

I think you missed my point. I think all of us agree with you and what you're saying. Indeed, many of the S meter implementations are really Vu meters, not S meters, which is kinda my point: The ?BITX is such a fun rig, does an "accurate" S meter actually augment the enjoyment we should get from it? To me, the answer is: "No!" It's nice to have an S meter, but I think we all know it's both subjective and relative and neither of those factors should detract from the fun. So, if you don't have an S meter, use your ears. If you do have one that works through the audio chain, use that. If yours work through the IF chain, good for you, but you're still going to give me a 59 during the next contest...aren't you? Who cares...as long as you can hear me!

Jack, W8TEE

On Monday, July 23, 2018, 11:11:51 PM EDT, Mark M <junquemaile@...> wrote:


Guys...I understand that S meters are mostly subjective and it's
certainly not a necessity but everything else works so well and the
capability is there so why not try to get it to work? I'm not looking
for mil-spec accuracy, I was just wondering if what I see is typical or
if it indicates a problem somewhere in the AGC board. I don't think I've
ever had a qrp rig with an S meter so it's kind of icing on the cake.

And like Jack, I've noticed that for some reason, during contests they
all seem to indicate S9.? ;)

Anyway, thanks for the inputs...

Mark? AA7TA




 

Correction to my previous post:
The dynamic range of this S-Meter will be greater than just the range of the diode detector,
since it is throttling back the RF gain control as the audio level rises.


On Mon, Jul 23, 2018 at 06:51 PM, Jerry Gaffke wrote:
An S-Meter using the diode detector of the AGC board will not have much range.
Given the two 1n4148 diode drops involved and 5v max into the Nano, I'd expect a range of under 3 S-Units.
At 6dB per S-unit, that's a voltage ratio of? ?10**(3*6/20) = 7.94.??
Use a log amp like the AD8307 if you really want an S-meter.?


 

Mark; Yes, those appear to be typical. The MOSFETs begin conduction at around a volt and a half.
I simply used a calibrated generator as a source to establish my S meter readings. Those values will differ slightly from unit to unit.
To those that have doubts about the attenuation I would advise making some measurements. At 3 volt bias there should be more than 50 dB attenuation at 7 MHz. Does the manual gain control work? If not, evaluate your RF grounding at the attenuator location for a start.
I agree, S meter readings are mostly ornamental. 73, Don


 

When you connect the S-meter to J1, don't forget that you still have to add a jumper to set the AGC "hold" time by "loading" C4. I expect the S-meter analog input to be a very small load

You might try a little different load on C4 by installing a different resistor jumpers on J2,J3,J4,or J5.? What effect does that have on the readings ? J5 has the highest 1.9M load and therefore the longest "hold" time.

73 Kees K5BCQ?