UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 31st May 2021, 2:21 pm   #1
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default RF preamp circuit selection

Local noise in far-away places makes travelling to do amateur radio rather hit and miss. Some places have noise levels of S5 or worse on HF. If/when we get over the Covid pandemic there are several places I'd like to visit with a radio but I'd like to find out in advance if reception will be satisfactory. Hence I've made a noise receiver for 14MHz costing less than £20 which I hope to courier to the landlord at a distant B&B and request a noise level measurement. If the site is hopelessly noisy I will just write off the receiver and try another destination.

The receiver topology I'm using is a preselector with about 1KHz bandwidth and 2dB loss at 14MHz made from 3 crystals, feeding an NE602 type crystal controlled mixer to produce a 10KHz audio signal that is amplifier by a NE5534 opamp and fed to an AD8310 log detector that directly drives an analogue meter calibrated in S-units. It works OK but is a bit noisy, registering S1 to S2 (around -120dBm) with the input terminated in 50 ohms.

So, after all that background waffle, here's my question. What circuit design and active component choice does the panel recommend for an amplifier to go between the preselector and the mixer and drop the noise floor by 6 to 10dB? The amplifier needs to present a decent 50 ohm input at 14MHz to terminate the preselector and it will benefit from a reasonable single-ended 50 ohm load presented by the mixer input tuned circuit.

I'm guessing a 10-12dB RF amplifier with a reasonably low noise figure and a 50 ohm input that draws maybe 10mA or less from a PP3 battery is "all" that is required? What should I use? I've read resistive feedback bipolar transistor amplifiers are noise but the David Norton "lossless feedback" circuit is much quieter (why?), though by how much I don't know. And anyway, how is the noise figure of this (or other) circuit topology predicted? General purpose transistors aren't specified for noise figure, or if they are it's at 1KHz rather than HF (2N3904, 2N2222, etc), and "microwave" transistors are unsurprisingly specified at hundreds of MHz and above. The .noise facility in LTspice seems to tell fibs, presumably because the transistor models are inaccurate.

A cookbook answer would be OK but literature references would help too, so long as they don't go mad with the maths.

Thanks, Alan G3XAQ
Alan_G3XAQ is offline  
Old 31st May 2021, 3:25 pm   #2
Guest
Guest
 
Posts: n/a
Default Re: RF preamp circuit selection

I am wondering if there are any "web SDRs" in the locale?
 
Old 31st May 2021, 3:31 pm   #3
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

Sadly no. I'm talking about Kampala, Accra, Kigali and so on. And it's very local noise that can cripple activity: SMPS, VDSL, and so on.

73, Alan
Alan_G3XAQ is offline  
Old 31st May 2021, 4:21 pm   #4
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: RF preamp circuit selection

Alan's checking for hotels/accommodation full of the nastiest LED lamps and SMPS wall warts.

In a normal feedback or degenerated amplifier, the gain is set by resistors acting as potentiometers. All resistors are noisy at room temp. So the transformer feedback idea is that the feedback potentiometer is replaced by an auto transformer. It's less noisy AND the other idea of using extra-low resistor values would waste output power.

Norton's patents are interesting. Q-bit went a bit further, but maybe Chris Trask is worth a google.

You shouldn't need such low noise figure. On HF even far from the nearest shonky switcher, you'll still hit galactic noise limits.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 31st May 2021, 4:40 pm   #5
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

Yes, I agree I don't need a super low NF on HF, but it would be handy of my prototype measurement receiver had at least a reasonable NF. It is registering about -120dBm internal noise in a roughly 1KHz front end filter bandwidth. That's -150dBm/Hz in 50 ohms or 23dB NF, which is poor by anyone's standards. The 10KHz stages after the mixer have an approximate bandwidth of 1KHz, maybe 2KHz. Data sheets for the NE602, NE5534 and AD8310 all suggest performance should be better. I do wonder if the mixer has much worse that the 5.5dB NF datasheet value when mixing down to 10KHz? Or perhaps I have other problems in my design.

Alan
Alan_G3XAQ is offline  
Old 31st May 2021, 7:32 pm   #6
Restoration73
Nonode
 
Restoration73's Avatar
 
Join Date: Jun 2013
Location: Surbiton, SW London, UK.
Posts: 2,801
Default Re: RF preamp circuit selection

I would always use a JFET or MOSFET for the application you describe.
Restoration73 is offline  
Old 31st May 2021, 9:58 pm   #7
jimmc101
Heptode
 
jimmc101's Avatar
 
Join Date: Mar 2009
Location: Portsmouth, Hampshire, UK.
Posts: 674
Default Re: RF preamp circuit selection

With a such a low frequency IF (10kHz) noise from the internal bias source of the NE602 can be a problem.
If you're running single ended, use a decoupling capacitors of at least 1uF on pins 2 & 8.
Your tuned circuit should ensure that pin1 also 'sees' a low impedance at 10kHz.
This may reduce the NF sufficiently to avoid the need for a pre-amp

Jim
jimmc101 is offline  
Old 31st May 2021, 10:48 pm   #8
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

My mixer input is balanced to pins 1+2, 3 turn 50 ohm single ended link from preselector and 27 turns plus 75pF trimmer on T50-6 micrometals core to pins 1+2. I even went so far as to add a gimmick capacitor on one side to get accurate amplitude and phase balance.

The DC feed to pin 8 is decoupled with 10uF plus 10nF.

Output is taken from pins 4+5 differentially to the NE5534 opamp.

Someone on another forum suggested disabling the NE602 (remove the DC supply to pin 8 but not the caps?) to confirm the excess noise is coming from the mixer and not from maladies later in the signal chain. I'd do that in the morning.

73, Alan
Alan_G3XAQ is offline  
Old 1st Jun 2021, 8:52 pm   #9
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
Default Re: RF preamp circuit selection

With such a low 'IF' you run into the area where the noise-sidebands of the local-oscillator become a serious issue! This was always a problem back in the days when I played-about with direct-conversion receivers using MD108/SBL1 diode-quads.

The problem being that a proper 'switching mixer' needed a really rather muscular LO drive - providing the gain needed to get several volts of LO drive into a 50-Ohm load invariably meant amplifying the noise-sidebands of the LO to the point where they approached the sub-microvolt signal you were trying to receive...

What is your proposed 'local oscillator'?? if it's anything synthesized then there's an old adage that "a synthesizer generates every frequency within its designed range, usually more than once at any one time".
G6Tanuki is offline  
Old 1st Jun 2021, 9:40 pm   #10
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

As I said, the mixer is an NE602 Gilbert cell using the on-chip LO, which is crystal controlled. I would hope that LO noise nearly 10KHz from the centre of the 1KHz wide preselector passband would not be an issue, but who knows?

I have some more measurements. If I disable the mixer by removing the DC supply to the chip (but leave the decoupling in place) the output from the log detector drops 14dB, more or less to the datasheet noise floor of the AD8310 log detector. Disabling the LO with a big capacitor across the crystal drops the output 8.3dB, as does shorting the link winding on the mixer input tuned circuit.

I am at a loss to explain why the mixer seems to be so noisy. Overall I can just detect a -120dBm carrier, suggesting the noise figure is a huge 23dB. The crystal filter preselector loses 2dB so the mixer has an apparent NF of 21dB. Comments on one of the QRP forums suggests a NE602 direct conversion receiver is very quiet and sensitive and "feels like" it achieves the datasheet 5dB noise figure.

I don't have any good ideas what is happening and what to try next.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 1st Jun 2021, 11:42 pm   #11
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: RF preamp circuit selection

You can get a 3dB rise in noise from the wide-open image band. With such a low IF you can run into phase noise problems.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 2nd Jun 2021, 8:59 am   #12
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

Yes, I've realised that it being a DSB receiver means the noise figure at best will be 3dB worse than the SSB NF quoted in the mixer datasheet.

I would hope the crystal local oscillator will be "quiet" but your hint that the image noise goes from DC to daylight makes me wonder if my audio stage that is mildly peaked at 10KHz is inadequate? Simulations say the response is 30dB down at 1KHz and 60dB down at 100Hz. The peak is pretty broad: -3dB at 7KHz and 35KHz. The log detector is broadband and will respond to any noise presented to it by the opamp.

If the preselector is 1KHz wide and the audio is 30KHz wide does it mean the overall noise figure is degraded by 14.7dB? Might this be the root of my dreadful observed 23dB NF? I'm getting a bad feeling about this...

73, Alan G3XAQ

Last edited by Alan_G3XAQ; 2nd Jun 2021 at 9:00 am. Reason: typo
Alan_G3XAQ is offline  
Old 2nd Jun 2021, 1:30 pm   #13
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

The consensus on other forums agrees with David. I'm pressed for space inside the box but I think I can turn my NE5534 audio stage into a multiple feedback bandpass filter with 2KHz bandwidth at 10KHz (Q=5). Any tighter makes me uneasy about transient response. It should be good enough to drop the noise floor below -130dBm.

Time to switch the soldering iron on...

Alan
Alan_G3XAQ is offline  
Old 2nd Jun 2021, 1:43 pm   #14
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF preamp circuit selection

I've been stung here in the past by wideband noise effects when designing a tiny low power telemetry receiver using a wideband limiter chip at the IF.

If you want to achive the expected performance you have to think in terms of generating a deliberately raised (narrow!) noise pedestal and you have to preserve this thin slice of noise to always have a noise power that dominates/defines the whole system. Otherwise the IF chip will detect/limit to the total noise power at its input and this could be over a very wide bandwidth.

I haven't done much with the SA602 or SA612 since the 1990s but Philips did produce lots of app notes describing some of its limitations. Jim is correct in that this chip needs to be decoupled correctly at the input at LF if you want to use it with an IF at LF. Otherwise it will inject its own excess LF noise into the mixer section and you will end up with more noise than you expect. However, I'm not sure this is the only cause of your issues as you need to consider what bandwidth the noise pedestal is you are generating ahead of the logamp.

I'm guessing the SA602 will produce 17dB conversion gain and you will need something to define a noise pedestal bandwidth after this. Your passive crystal filter is too near the front end to achieve any kind of raised noise pedestal above thermal noise.

I would suggest XTAL_FILTER >> SA602 >>BJT stage >> AF bandpass filter >> logamp

Hopefully the conversion gain of the SA602 plus the gain of the BJT stage will produce your noise pedestal when viewed at the output of the (active?) BPF and this narrow pedestal of noise power needs to be powerful enough to offset/dominate the wideband contribution to noise at the (400MHz) logamp input.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Jun 2021, 2:04 pm   #15
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF preamp circuit selection

If you can get hold of a SA615/SA605 this has the same front end mixer and it also has a limiter that has a fairly good (90dB) RSSI response in terms of dB/mV. In your case the RSSI response doesn't have to be as accurate as the AD8310 chip because you can simply use a lookup correction based on a calibration using a decent sig gen.

The above system would hopefully be a lot cheaper and would run at low power and the limiter has less bandwidth and it can be further noise bandwidth limited at the mid point of the IF amp/limiter stages in the SA605 to just a few kHz if you desire.

https://www.nxp.com/products/radio-f...f-system:SA605
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Jun 2021, 3:09 pm   #16
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

Hello Jeremy. I can see where you are headed but I don't need the performance you can achieve with the fancy chip or extra amplifiers. Africa is always fairly noisy (distant thunderstorms, etc) and I've never seen anything quieter than S1 in 400Hz on my Elecraft K3. So an MDS around -130dBm in 1KHz or 2KHz will be ample, corresponding to a noise figure around 10dB. I *think* if I convert my NE5534 audio amp into a 2KHz wide multiple feedback BPF then I might get to where I want to be without major surgery to the existing prototype.

I only really need performance good enough to answer the question: "Is the receiver seeing a noise level above or below S4?" My existing unit can just about answer than but I'd like a bit more headroom and I hated not knowing why it was so noisy.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 2nd Jun 2021, 3:29 pm   #17
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF preamp circuit selection

Ok, it sounds like you are close to what you need. With a wideband logamp you will also have to be wary of any sneaky LO (and LO harmonic) leakage into the 400MHz logamp. Also ground loop currents and LF pickup on a test bench can cause some problems and headscratching too.

One of the first receivers I designed was back in the early 1990s and it was the little low power matchbox sized telemetry receiver. This had to work all around the world and could be buried in ice somewhere cold and one particular spec requirement was for it to work without overheating or drifting when buried and hidden in hot sand in an unfriendly part of the world.

This was the design that caught me out with the issues you are seeing. I didn't use an SA602 mixer in this design but I did have similar noise bandwidth issues that were happening after the main IF crystal filter. This reduced the sensitivity well below what was expected and I had to include another bandwidth defining filter further down the chain. This is similar to what you are about to do now but I had to do it at the IF frequency in my case.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Jun 2021, 8:05 pm   #18
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF preamp circuit selection

If it helps I dug out my stash of old eval boards for the SA602/612. These mixer chips are probably 25-30 years old and have seen a lot of tinkering in terms of the configuration over the years and the boards are a bit scruffy now.

All three boards are designed for VHF input and about a 10.7MHz IF. However, I added a dual DC coupled BJT based baseband stage at the output of one of the SA602 chips.

This board has a 50MHz balanced input using a T37-6 based transformer and it looks like it has 3T on the primary and 11T +11T making up 22T for the centre tapped secondary. This feeds to pins 1 and 2 of the SA602 and there is a small tuning cap there and this is probably a few pF in value.

To get low noise performance down at a 10kHz IF I had to boost the ceramic shunt cap at the centre tap to a big SMD tant cap. The centre tap of the balanced secondary winding has to be AC coupled because there is a DC bias voltage here.

I just measured the noise figure using an accurate sig gen and a spectrum analyser and it showed just under 7dB noise figure for the 50MHz downconversion to a 10kHz IF.
Some of the noise contribution to this 7dB figure will be nearby image noise due to the low IF.

With the original ceramic cap at the centre tap ( it was probably 10nF or 100nF) the noise figure was definitely worse by several dB.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Jun 2021, 8:13 pm   #19
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF preamp circuit selection

I should have measured the noise figure with the original 10.7MHz IF before I did all this but I think it was just under 5dB the last time I played with it. I'll be putting it back the way it was later so I can repeat this test when I do this.

Note: the last time I measured the noise figure was with the PSA spectrum analyser and an HP 346 noise source as it can calibrate and measure the noise figure of a mixer with this setup. This gives really low uncertainty for the measurement. Much better than the method I just used with the sig gen. However, I can't repeat this with the PSA analyser and the noise source if the IF is down at 10kHz.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Jun 2021, 8:18 pm   #20
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: RF preamp circuit selection

(resting my eyes, half way through soldering the parts for the 10KHz BPF).

Is a centre tapped input all that important? At 14MHz I'm seeing reasonable balance with just a 2pF gimmick cap on one side of the winding but with it otherwise floating (fig 4b on the NXP SA602A datasheet).

Alan
Alan_G3XAQ is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 8:57 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.