The AM Forum
March 28, 2024, 08:02:37 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1] 2   Go Down
  Print  
Author Topic: Boat Anchor Receivers - AM Audio S/N Measurements  (Read 59137 times)
0 Members and 1 Guest are viewing this topic.
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« on: October 13, 2008, 09:45:21 PM »

In another receiver dynamic range thread there was some interest in measuring boat anchor receivers for recovered AM audio signal to noise.   

The receiver under test is injected with a 1000 uV rf signal, 30% amplitude modulated at 1 kHz, and the resulting signal to noise is measured. This level is sufficient to fully 'quiet' the receiver and this test is a good indication of the performance of the i-f, detector and audio stages of the receiver. The higher the value the better.

For these measurements an HP-8640B was the signal source and an HP-339A Distortion Analyzer was used as the calibrated voltmeter in order to take advantage of its 400 Hz high pass filter. The high pass filter effectively removed 60 and 120 Hz receiver hum from the measurement.   


Receiver    Audio S/N Ratio
R390A      41 dB 
75A4      48 dB   
GPR 90      45 dB   
HRO 60      52 dB 
SX-28A      52 dB   
Drake 2-A      33 dB   

The audio S/N information has been added to the dynamic range measurements page at 

http://www.w1vd.com/BAreceivertest.html

A possible additional test occurred to me while using the HP-339A and HP-8640B combo. Output from the audio generator section of the HP-339A could be used to amplitude modulate the HP-8640B generator and the receiver recovered audio applied to the distortion analyzer section of the HP-339A. One could then chart recovered audio distortion vs. modulation level...say every 10% up to 100% modulation.

Not sure about the absolute accuracy but it may be good on a relative scale. Thoughts?       

Logged

'Tnx Fer the Dope OM'.
Ian VK3KRI
Guest
« Reply #1 on: October 14, 2008, 07:14:33 AM »

Interesting numbers.  Is that measurement the ratio of 1000uV signal with 30% 1khz tone vs 1000uV signal with no modulation, or  1000uV signal with 30% 1khz tone vs the same think with the tone notched by the distortion meter ?
If its the second you're picking up detector distortion as well.   

When I started playing with my TX audio quality, I quickly found that I didn't have any receiver close in noise or distortion that to what I could get out of the transmitter. I guess I really should build a low distortion detector one day..
                                                                                              Ian VK3KRI   
Logged
w3jn
Johnny Novice
Administrator
Member

Offline Offline

Posts: 4620



« Reply #2 on: October 14, 2008, 12:31:45 PM »

Very, very interesting, Jay! 

The distortion measurement would be extremely useful.  Many receivers sound OK up to about the 70% modulation level, then really fall to pieces.  Many Hallicrafters receivers are in this category - SX-101, SX-100, etc., due to the poor ANL design.  Particularly interesting would be receivers with mechanical filters vs those with L/C vs those with crystal filters. 

A couple of the best-sounding Am detectors are those in the WJ-8716 and Drake R-7.  No fuzziness even on heavy peaks.  I thought my GPR-90 sounded decent until I used it with a Sherwood sync detector - I've now pretty much given up on the internal detector in the '90.
Logged

FCC:  "The record is devoid of a demonstrated nexus between Morse code proficiency and on-the-air conduct."
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #3 on: October 15, 2008, 01:26:27 PM »

Jay,

Sounds good on the distortion data also.  SInce we are interested in good AM receivers, all this audio data is important too.

Thanks for your efforts.
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
Jeff W9GY
Member

Offline Offline

Posts: 257



« Reply #4 on: October 15, 2008, 05:31:58 PM »

I would think it useful to perform this test at 10 uV and 100 uV  signal levels, too.  This would give a good indication of the the receiver's overall quieting characteristic as the signal increases. I have a feeling some receivers will produce a better S/N at lower inputs than others who have good numbers at 1000 uV.
Logged

Jeff  W9GY Calumet, Michigan
(Copper Country)
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« Reply #5 on: October 16, 2008, 08:45:21 PM »

Had a chance to try the HP-339A / HP-8640B setup to measure receiver AM detector and audio stage distortion. The low distortion generator output from the HP-339A was used to modulate the HP-8640B. Spec for the AM modulator in the 8640B is <1% distortion and measurements of the 8640B modulated output on a spectrum analyzer confirm this to be the case - second harmonic is 40 dB or better down and higher order harmonics are considerably further down. One could, in theory, measure THD down to some level approaching -40 dB.

Since the Drake 2-A was still on the bench it was used for the initial test. The audio output from the 2-A was connected to the analyzer section of the HP-339A. The auto gain set and auto null features appeared to operate correctly. Although not knowing what to expect, it was a little surprising to see a distortion level of only -29 dB at 30% modulation. Could this really be?

As a double check the receiver output was also connected to a Delta 44 sound card / computer and an FFT program (Spectrum Laboratory) was used to view the receiver audio output. The FFT spectral displays for different modulation levels are shown below. The calculated THD, taking into account all the major harmonics, matches up to that displayed by the HP-339A within 1 dB so it seems the test setup is working properly.

So...a couple questions...

Since I'd rather not make more measurements than necessary, are measurements at 30, 50, 70, 90 and 100% necessary or appropriate? At what audio output power level should they be made...1/4, 1/2 or full rated power? Don't imagine most folks have their receivers cranked to the max. Are the spectral displays better than just a simple THD figure? Are there other forms of distortion we should be looking for or will THD be the significant form of distortion?

Some of these questions may be difficult to answer until a few more receivers are run through the test. May not get a chance to do that until next week so thought I'd throw this open for discussion in the meantime.

                     


* 2A 30%.gif (41.91 KB, 788x513 - viewed 1922 times.)

* 2A 50%.gif (42.04 KB, 788x513 - viewed 1916 times.)

* 2A 70%.gif (42.32 KB, 788x513 - viewed 1915 times.)
Logged

'Tnx Fer the Dope OM'.
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« Reply #6 on: October 16, 2008, 08:46:30 PM »

And the last two plots...


* 2A 90%.gif (43 KB, 788x513 - viewed 1933 times.)

* 2A 100%.gif (42.4 KB, 788x513 - viewed 1900 times.)
Logged

'Tnx Fer the Dope OM'.
w8khk
Member

Online Online

Posts: 1200


This ham got his ticket the old fashioned way.


WWW
« Reply #7 on: October 16, 2008, 09:06:00 PM »

I would think we would want to see the distortion measurements performed at a low enough audio amplifier power level such that the results indicate the distortion of the just the detector at the various modulation levels.  We know the communications receivers are noted for poor quality audio output, especially when pushed to high levels, and we often use external "hi-fi" amplifiers to get decent audio.  If the measurements are a mix of detector and audio amplifier distortion, they would be rather meaningless, in my opinion.  To minimize the number of tests, run the audio low enough to see just detector distortion.  Just my two cents.....
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #8 on: October 16, 2008, 09:22:10 PM »

Jay,

I just checked one of my TS-430's and my R-7.  I didn't have a HP8640B years ago when I first tested the receivers, I got:

Measured 10/16/08 with my HP8640B:

Speaker output driven to 1 volt r.m.s.  Checked with scope to ensure no clipping.

stock Kenwood TS-430 = 41 dB S/N
Drake R-7 (some modifications) 46 dB S/N.

I have an Excel sheet I can send you that calculates the % THD.  Handles plugging in the 2nd, 3rd, 4th and 5th harmonics xx dB down from fundamental.  Can leave highest harmonics defeated by plugging in "200".

Eyeballing your 2A 70% graph and plugging in 2nd, 3rd and 4th harmonics, I got 6.155 % THD.

90 % graph; 11.8 % THD,

100 % graph; 13.1 % THD,

* 30 % graph; 3.4 % THD.
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #9 on: October 16, 2008, 09:30:52 PM »

Jay,
I just checked the audio specs for the RA 6830. 600 ohm balanced line out is 2% at 0 dB and high level spec is 5%. So 10 % would be 20 dB and 5% would be 26 dB. Also 2% is what 34 dB.
just some numbers to compare with your findings.
Logged
w3jn
Johnny Novice
Administrator
Member

Offline Offline

Posts: 4620



« Reply #10 on: October 17, 2008, 07:40:04 AM »

Jay, I would think that a low output would be good so that, as Tom offered, the detector distortion is being measured more than the distortion inherent in the audio stages.  Best would be to tap at the volume control, as you will have certain RXs with shaped audio (both in the low level stages andoutput xformer) that will attenuate higher harmonics and therefore present a lower (yet invalid) distortion reading.

I also think that testing at 50% and 80-90% mod lvels should be fine.  Most should be pretty good at 50%; where they fall apart is at the higher modulation indexes.  Or so it seems to me. 
Logged

FCC:  "The record is devoid of a demonstrated nexus between Morse code proficiency and on-the-air conduct."
Rob K2CU
Contributing
Member
*
Offline Offline

Posts: 346


« Reply #11 on: October 17, 2008, 09:49:49 AM »

You need to decide what you are trying to measure in the first place, as a means to compare different receivers. The distortion and noise that comes out of an AM receiver comes from four sources:  The input signal noise and distortion; RF/mixer/IF stages IMD and noise; the AM detection process; and the audio output stage. A purest approach would be to consider the whole radio and look at the speaker terminal outputs. With the exception of a few radios where they went out of their way to design a crappy audio output stage, the predominant source of distortion and noise in an AM radio will probably be the detector stage itself.

Simplistically, there is an idea that a diode detector works by erasing the bottom half of an AM modulated carrier and then putting it through a low pass filter to sooth out the result. In actuality, diodes are quite non linear devidces. The output of the last IF stage is impressed upon a diode in series with a "load" resistor. The current through the diode follows either of two basic I-V equations depending upon whether it is a vacuum tube diode or a bipolar semiconductor diode.  For a tube, the equation is approximately:

I = k V ^ 3/2 where k is some constant.

For a semicondutcor it is:

I = e ^ k V  where this k is some other constant.

the exponential can be represented by the infinite series:

e^x = 1 + x + 1/2 (x^2) + 1/6 (x^3) + 1/24 (x^4) +......

When our input is a sum of three sin or cos signals, the result is a slew of intermodulation harmonic distortion products for either type of detector.

Even noise is worse in an AM "envelope" detector. To show that, just consider the difference in noise with no antenna connected, between SSB mode and AM mode. Also, you can "hear" and the unmodulated carrier presence due to the detected nopise rise and even tuen it in.

It might be interesting to run some tests to not only compare entire radios, but also to devise a test to bring out the IF output and use a low distortion SDR receiver to look at RF circuitry generated harmonic distortion, the detector output, and the speaker output. Arbitrarily set some fixed reference points, like 50 uV carrier level, 80% modulation, and 100 mW audio output into the speaker.

It has been my experience that the bulk of the problem has been the detector. And it usually is the problem that at higher levels of modulation, the detector falls apart when the modulation presses the carrier towards zero level. Rarely does the modulation peak cause a problem by hitting the roof.

But, you might want to step back and consider what you are trying to accomplish in the first place. That AM sound that we have known from our early days in radio when AM was king, was always loaded with rich harmonincs that gave it a warm, full bodied sound in your shack. My whole venture into imroving my R4-C and TR-7 started when I was listening to the grey haired net one night and decided to fire up my old HE-30. I could not believe how beautiful the sound was coming out of that radio compared to the newer and way more expensive rigs. Well, not too surpriseing as the HE-30 was designed as a AM SW receiver that happened to have a BFO added to it for CW/SSB. The other two were SSB receivers first, with AM added as almost an afterthought.

Perhaps I shoud l run some tests measuring single tone SSB distortion, DSB AM distortion, and SSB AM distortion. Now to find the time.
Logged
KM1H
Contributing
Member
*
Offline Offline

Posts: 3519



« Reply #12 on: October 17, 2008, 11:19:18 AM »

Jay, I had a Drake 2B with 2BQ around 1963. The audio was so poor I quickly got rid of it. No measurements made.

In the mid 80's I had a couple of R-4C's that were also known for distorted audio. I changed to the LM something or other IC that was recommended by Sherwood and it was a huge improvement. Again no measurements taken so its strictly subjective but it became a pleasure to use even digging for 160M DX.

Carl
KM1H
Logged
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #13 on: October 19, 2008, 07:36:00 PM »

I took a stab at the audio distortion measurements this afternoon with a Kenwood TS-430 and a Drake R-7.  Although it would seem that this kind of measurements should be duck soup, I question the results some. 

HP8640B generator, 1000 Hz AM modulation, 1000 uV. of r.f. at 3838 kHz. 
HP331A connected to speaker terminals - output 1 V. r.m.s.

                  TS-430    R-7
30 % mod.      1.3       1.2 % THD
50% mod.       1.5       1.0
70 % mod.      1.55      0.75
90 % mod.      1.6        0.56
"100" % mod.   7.9       0.7 % THD

The TS-430 AM detector is a diode.  6 kHz. BW.

The R-7 AM detector is a product detector/synchronous detector.
* AGC set to SLOW, NB and r.f. pre-amp are off.  6 kHz. BW.
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #14 on: October 19, 2008, 09:06:57 PM »

Tom,
I wonder if you could scan the R7 AM detector schematic and send me a copy. I guess the R7 is different than the TR7. TNX Frank
Logged
Rob K2CU
Contributing
Member
*
Offline Offline

Posts: 346


« Reply #15 on: October 20, 2008, 07:53:23 AM »

Hi Frank,

I think I read somewhere that the AM detector in the R-7 is similar to that used in the Racal 6790. The signal is split into two paths, both sent to the product detector, except that one goes through a limiter first. As you know, the TR-7 uses an unimaginative diode envelope detector.  I don't recall the signal level presented to the LM1496 product detector on the Rf input terminals. If high enough, one could consider running it into the LO input as well, inplace of the diabled BFo when in AM mode. A couple of PIN diodes would be needed to do this switching when AM mode is selected, and the Audio would always from from the 1496. I haven't tried it, but one could always pull the IF/detector/audio board and JS the connection to try it out. The LO input o fthe 1496 has a built in limiter, so it migth work fairly well and be a simple mod to do.

A more complex solution would be to also run the IF through the CW filter and then amplifiy it with a single FM IF amplifier IC and run that signal into the LO input of the 1496. the CW fillter would strip off most sideband artifacts from the carrier before amplifying and limiting it. Call it a "Wazoodyne" detector. ha ha!  There would probably be a need for phase adjustment to compensate for the different paths and delay. The narrow CW filter would make the carrier path at least ten times more sensitive than the signal path.

Most of these ideas were tried out in various forms and even patented back in the early thirties, during the developmental haydays of AM radio. thet didn't have the narrow filters we have today, but they often used regernerative oscialtors that would lock to the carrier drive.

Logged
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« Reply #16 on: October 20, 2008, 04:39:29 PM »

Thanks all, for the comments.

The previous receiver tests to measure MDS, blocking and two-tone dynamic range and AM audio S/N measurements are straightforward and the results match well with the experience of actually using the receiver connected to an antenna.

As far as AM distortion measurements, agree that high levels of harmonics give that 'warm' and 'full' AM sound. Some audiophiles have proposed that relatively high levels of harmonics are not objectionable...as long as they decay in an orderly fashion as their harmonic number increases. What appeared to be particularly objectionable (to them) was excessive levels of higher order harmonics. This makes me think a spectrum plot for the AM distortion may be appropriate rather than just a THD 'figure of merit'. May not get back to measuring distortion for another week or so not sure what I'll find with other receivers.     

The intention of these receiver tests is to document performance - not necessarily to devise ways to improve performance. Many of the receivers going forward will not be available for modification.

So far we've got:

1) Measure at low audio output levels to presumably see just detector distortion

2) Measure two or three modulation levels (50%, 80%, 100%)?

3) Include spectrum plot

Rob...agree that it would be interesting to look at the rf input to the detector for harmonic content and will do that as time permits. BTW, I use your precision rectifier detector on the 455 kHz output from my R-390A...along with a sync detector and SoftRock / PowerSDR setup. Output from the three detectors are selected at the Hafler DH-110 preamplifier for quick a-b-c comparison. Your circuit works very well.

 

                           

Logged

'Tnx Fer the Dope OM'.
w4bfs
W4 Beans For Supper
Contributing
Member
*
Offline Offline

Posts: 1433


more inpoot often yields more outpoot


« Reply #17 on: October 20, 2008, 07:01:51 PM »

hi Jay ... this is some really good work that you've done here ... about distortion, the human ear is relatively insensitive to even order harmonics but is quite sensitive to odd order harmonics ...second harmonic distortion is perceived as 'warmth' or 'fullness' where 3rd and 5th harmonic distortions are perceived as 'rough, harsh, or grating' ...odd stuff hearing ...couple this with the Fletcher-Munsen syndrome (amplitude/frequency/spl variances) and its no wonder we have such differences of opinions as to what sounds 'good' ....73 ...John
Logged

Beefus

O would some power the gift give us
to see ourselves as others see us.
It would from many blunders free us.         Robert Burns
Ian VK3KRI
Guest
« Reply #18 on: October 21, 2008, 06:29:11 AM »

hi Jay ... this is some really good work that you've done here ... about distortion, the human ear is relatively insensitive to even order harmonics but is quite sensitive to odd order harmonics ...second harmonic distortion is perceived as 'warmth' or 'fullness' where 3rd and 5th harmonic distortions are perceived as 'rough, harsh, or grating' ...odd stuff hearing ...couple this with the Fletcher-Munsen syndrome (amplitude/frequency/spl variances) and its no wonder we have such differences of opinions as to what sounds 'good' ....73 ...John

Here's a link to some files I created to actually hear the difference for myself...
http://www.smoke.com.au/~ic/audio/dist_samples.html
Ian vk3kri
Logged
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #19 on: October 22, 2008, 01:27:39 PM »

Frank,

I emailed you a doctored camera jpeg of the detector area last night.  The R-7 has an additional conversion to a third i.f. of 50 kHz to support the op amp tunable notch filter and the syncronous detector, among other differences:
Internal ac supply, switchable r.f. pre-amp, half-octave r.f. bandpass filters, antenna splitter, antenna switching.
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #20 on: October 22, 2008, 01:50:51 PM »

TNX Tom,
I sent you a copy of the Racal set up. The difference is the Racal near zero signal level still generates a carrier but the phase noise gets wider. The drake would be just be IF noise. I've often considered putting a 455 KHz pll in series with the 1357 and 1496 using an XOR gate as a phase detector so no signal throws the carrier to 455 khz. (XOR at 50% duty cycle)
This detector beats SDR when conditions are bad every time. I bet a real sync detector will also beat the bit bashers......only a matter of time though.
Logged
Steve - WB3HUZ
Guest
« Reply #21 on: October 22, 2008, 02:13:58 PM »

Interesting. The higher freq harmonics (2205 and up) stood out like a sore thumb, even at the 0.1% level. The lower ones were noticable at the 1% level but barely, if at all, at the 0.1% level. This was listening on a crappy PC with the sound card on the motherboard and cheapie speakers. I'll try it again at home on the Mac, with headphones.


hi Jay ... this is some really good work that you've done here ... about distortion, the human ear is relatively insensitive to even order harmonics but is quite sensitive to odd order harmonics ...second harmonic distortion is perceived as 'warmth' or 'fullness' where 3rd and 5th harmonic distortions are perceived as 'rough, harsh, or grating' ...odd stuff hearing ...couple this with the Fletcher-Munsen syndrome (amplitude/frequency/spl variances) and its no wonder we have such differences of opinions as to what sounds 'good' ....73 ...John

Here's a link to some files I created to actually hear the difference for myself...
http://www.smoke.com.au/~ic/audio/dist_samples.html
Ian vk3kri
Logged
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« Reply #22 on: October 24, 2008, 11:32:52 PM »

Had some time available tonight so decided to take a close look at the HRO-60 AM distortion characteristics using the HP-8640B/HP-339A combo. Once set up, the THD measurements went quickly so this test could be easily applied to the BA receivers previously tested for MDS, dynamic range and audio S/N.

Results are as follows:

Receiver: National HRO-60

Mod  100 Hz  200 Hz  400 Hz  600 Hz   800 Hz  1 kHz   2 kHz
30%  -26 dB  -29 dB  -31 dB  -32 dB  -33 dB  -33 dB  -31 dB
50%  -23 dB  -28 dB  -32 dB  -34 dB  -34 dB  -35 dB  -35 dB
70%  -22 dB  -27 dB  -31 dB  -33 dB  -34 dB  -35 dB  -36 dB
90%  -21 dB  -26 dB  -30 dB  -32 dB  -32 dB  -33 dB  -33 dB
100%  -19 dB  -22 dB  -25 dB  -27 dB  -29 dB  -31 dB  -35 dB

Or, if you prefer the results in terms of %:

Mod  100 Hz  200 Hz  400 Hz  600 Hz   800 Hz  1 kHz   2 kHz
30%  5.0%  3.5%  2.8%  2.5%  2.2%  2.2%  2.8%
50%  7.1%  4.0%  2.5%  2.0%  2.0%  1.8%  1.8%
70%  7.9%  4.5%  2.8%  2.2%  2.0%  1.8%  1.6%
90%  8.9%  5.0%  3.2%  2.5%  2.5%  2.2%  2.2%
100%  11.2%  7.9%  5.6%  4.5%  3.5%  2.8%  1.8%

While making these measurements an eye was kept on a spectrum display looking for any unusual harmonic distributions...none were observed.

There may be more data points here than is necessary to adequately characterize receiver AM distortion performance...this may become obvious after a few more receivers are measured. 

     






Logged

'Tnx Fer the Dope OM'.
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #23 on: October 25, 2008, 12:54:41 PM »

Say you take an HRO 60 and rip it apart and replace all the resistors and caps with modern parts then replace all the wire with nice teflon insulated wire.
These old receivers with a high Q front end seem to be quite good.
Now take this high q front end and hang a good SDR RX after the tuned RF amp and I bet you get interesting performance. 
These old receivers seem to have very good MDS while dynamic range still quite good. Heck most of these receivers can handle a 10 dB pad at the input and still have good MDS giving an additional 10 dB of dynamic range.
Logged
Steve - WB3HUZ
Guest
« Reply #24 on: October 25, 2008, 03:23:18 PM »

Wow. The distortion levels at the lower frequencies are pretty bad. No wonder some people complain about mushy bass.
Logged
Pages: [1] 2   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.11 seconds with 19 queries.