EMRFD Message Archive 10812

Message Date From Subject
10812 2015-03-13 09:18:56 k1rf_digital_stev... Mizer front end question
Hello all. This is my first post. Allow me to introduce myself. I'm a retired electrical engineering manager who superved a digital design group. I have always had a dream of building a good performing HF communications receiver with the modern components of the day and, now that I gave the time, am proceeding with it.  I have a few questions for you, primarily for Ashar Farhan.  I am inspired by some ideas in the minima transceiver (see 

http://www.phonestack.com/farhan/minima.html) and also by the HF7070 receiver published in QEX. See:

http://www.arrl.org/files/file/QEX_Next_Issue/Jul-Aug_2013/QEX_7_13_Horrabin.pdf  I'm shooting for a balance between simplicity and good performance.  Now here's my questions:


1. A saw a mixer circuit in the VU2ESE folder for a KISS mixer/SI570/divide by 2 flop flop. Has that circuit been built and tested?  I like its simplicity


2.  I plan to use a homemade 6 pole crystal filter of the "Constant equiripple or QER GeUUR variety". I found some fundamental crystals at 29 cents each which are fundamental crystals at 36.86 mhz. I thought this might be useful for continuius general coverage but may be too lossy compared to 20 MHz crystals?  The IF would be a little too close to the upper RF frequency, making front end filtering difficult.


My head is spinning from looking at different mixewr termination techniques for best performance. (Hi dynamic range RF amps+ attenuators, diplexers, terminating the RF port at all frequencies as well as the IF port, hybrid connected dual crystal filters, etc).  I may be overthinking all the possibilities.  My current approach is to use a KISS mixer a la the Minima driving the 8 pole QER filter driving a PHA-11 amplifier chip with an attenuator pad.  The second mixer would also use a KISS mixer going down to 15 KHz, feeding a 24 bit audio A/D chip feeding a DSP IF.  For the 2 oscillators, I would use a silicon labs SI5338 for multiple outputs and low phase noise. I'd use divide by 2 flipflops on the oscillators for accurate 50% duty cycle,


Any thoughts or comments greatly appreciated.  Thank you.


"Digital Steve", K1RF


10813 2015-03-13 09:51:38 Thomas S. Knutsen Re: Mizer front end question
If your crystals are at 36.86MHz, there is NO WAY they can be fundamental. The physics don't allow making crystals this high reliable. 
There is a lot of datasheets who claim that they make fundamental crystals to several hundred MHz, but measurements will show them to be overtone crystals.  

I bet your crystals will be at 12.2MHz. A good choise of a IF in a HF transceiver. Measurements show that filters made with overtone crystals have less unwanted resonances and better stopband attenuation. A fine place to get out those old 27MHz crystals and put them to use!. 

If you don't have it, get EMRFD, read the chapter on transceivers (ch. 6) and do the receiver preformance calculations. 

73 de Thomas LA3PNA.

2015-03-13 13:42 GMT+01:00 sbdick@optonline.net [emrfd] <emrfd@yahoogroups.com>:
 

Hello all. This is my first post. Allow me to introduce myself. I'm a retired electrical engineering manager who superved a digital design group. I have always had a dream of building a good performing HF communications receiver with the modern components of the day and, now that I gave the time, am proceeding with it.  I have a few questions for you, primarily for Ashar Farhan.  I am inspired by some ideas in the minima transceiver (see 

http://www.phonestack.com/farhan/minima.html) and also by the HF7070 receiver published in QEX. See:

http://www.arrl.org/files/file/QEX_Next_Issue/Jul-Aug_2013/QEX_7_13_Horrabin.pdf  I'm shooting for a balance between simplicity and good performance.  Now here's my questions:


1. A saw a mixer circuit in the VU2ESE folder for a KISS mixer/SI570/divide by 2 flop flop. Has that circuit been built and tested?  I like its simplicity


2.  I plan to use a homemade 6 pole crystal filter of the "Constant equiripple or QER GeUUR variety". I found some fundamental crystals at 29 cents each which are fundamental crystals at 36.86 mhz. I thought this might be useful for continuius general coverage but may be too lossy compared to 20 MHz crystals?  The IF would be a little too close to the upper RF frequency, making front end filtering difficult.


My head is spinning from looking at different mixewr termination techniques for best performance. (Hi dynamic range RF amps+ attenuators, diplexers, terminating the RF port at all frequencies as well as the IF port, hybrid connected dual crystal filters, etc).  I may be overthinking all the possibilities.  My current approach is to use a KISS mixer a la the Minima driving the 8 pole QER filter driving a PHA-11 amplifier chip with an attenuator pad.  The second mixer would also use a KISS mixer going down to 15 KHz, feeding a 24 bit audio A/D chip feeding a DSP IF.  For the 2 oscillators, I would use a silicon labs SI5338 for multiple outputs and low phase noise. I'd use divide by 2 flipflops on the oscillators for accurate 50% duty cycle,


Any thoughts or comments greatly appreciated.  Thank you.


"Digital Steve", K1RF


10814 2015-03-13 10:03:35 vbifyz Re: Mizer front end question
Si5338 has high quality output drivers. When the output divider ratio is even, it should give a better 50% duty cycle than most of the discrete flops.
Two of its outputs can be frequency locked and phase offset by 90%. You can use it as a VFO for a zero IF receiver (mix directly from RF to DC/baseband), without any additional flops. There will be some I/Q imbalance because of the residual phase error and mixer component mismatch, but there are techniques to null it in the DSP/SDR part.
If low cost IS the goal, then Si5351 is the VFO part to use and zero IF is definitely the way to go.
Disclaimer: I am also a digital designer, RF is just my hobby.

73, Mike AF7KR
10815 2015-03-13 10:04:58 kb1gmx Re: Mizer front end question


1. A saw a mixer circuit in the VU2ESE folder for a KISS mixer/SI570/divide by 2 flop flop. Has that circuit been built and tested?  I like its simplicity


It has issues with LO feed through to the antenna.  This has been discussed on the minima

listserver.  (minima@freelists.org).  You can sign up and read history.


Minima is a evolving transceiver  design and there have been many suggestions on how to do it better or easier.  Among those better mixers and IF amplifiers.  As a receiver getting the if up and away from the received band is a desirable thing.


2.  I plan to use a homemade 6 pole crystal filter of the "Constant equiripple or QER GeUUR variety". I found some fundamental crystals at 29 cents each which are fundamental crystals at 36.86 mhz. I thought this might be useful for continuius general coverage but may be too lossy compared to 20 MHz crystals?  The IF would be a little too close to the upper RF frequency, making front end filtering difficult.


Actually 36.86mhz is good.  The biggest issue with minima was 20mhz is cost to both 17m and 15M.  It uses a trap to keep 20mhz out of the way but a trap at that frquency is both lossy to the over all input filter and not narrow enough to prevent interaction at 18 and 21 mhz.  There is no fixed filter loss number that makes for greatness other than no loss at all.  That means your always trading loss for filter performance such as bandwidth, stopband attenuation, and filter shape.  The filter loss can be fixed with IF gain.



My head is spinning from looking at different mixewr termination techniques for best performance. (Hi dynamic range RF amps+ attenuators, diplexers, terminating the RF port at all frequencies as well as the IF port, hybrid connected dual crystal filters, etc).  I may be overthinking all the possibilities.  My current approach is to use a KISS mixer a la the Minima driving the 8 pole QER filter driving a PHA-11 amplifier chip with an attenuator pad.  The second mixer would also use a KISS mixer going down to 15 KHz, feeding a 24 bit audio A/D chip feeding a DSP IF.  For the 2 oscillators, I would use a silicon labs SI5338 for multiple outputs and low phase noise. I'd use divide by 2 flipflops on the oscillators for accurate 50% duty cycle,


Are you reading EMRFD, if you don't have it get it. Seriously.  There is a lot of combined thinking in there.  It explains what the compromises are and how one might evaluate and balance them.  

You can over think it.  EMRFD, part of the books ideas what building everything to operate as 50 ohm modules, building  the module and then testing it for the performance provided
and maybe adjusting it or even replacing it.   so if you take all of the above and work as modules you can then see if they work as believed (and maybe modeled) and include them or build different.  Its a very good way to build classic analog radios.

The SI570 is currently lower phase noise than the 5338.  Though the 5338 is good.
There is a mixer using the FST3257 bus switches worth looking at as well.

FYI also check the R2pro and also the softrock list.  If your doing a IQ SDR system then crystal filters in the IF or even an IF are maybe not the best path .  The IQ SDR
can be simple down conversion to baseband and all the rest analog(at audio) or digtial processing.   That technique is very good and can be both very good and amazingly 
simple hardware. For existing hardware (kits) look at Softrock series 

Having built mixed mode circuits, embedded processing and RF gear I can say you have a lot of choices may are simpler than a classic dual conversion to a digital IF yet every bit
capable of high performance receive (and transmit).  That and newest high performance
cpus in the 32bit (ARM, MIPS, Intel) and larger realm allow serious processing power
some at seriously low dc power.  

I've built the Softrock Ensemble Transceiver as a means to test the waters fast
and easy and also work with the existing software.  There is a bare receiver too.

Look at Five Dash Inc. Your Source for SoftRock Radio Kits

 

  also the yahoo groups Softrock interest group.


Allison

10816 2015-03-13 10:05:26 Lasse Moell Re: Mizer front end question
Would not be so sure about fundamental frequencies... I have been reading about crystals where the center has been etched a lot thinnner that the outer part which acts as a frame. Hence resontates at higher frequencies but not as fragile as a "normal" crystal. There has been OT crystals at ~200 MHz and not 9th but more like 3d or 5th....

I am sure other who knows this stuff will tell us all :D

/Lasse SM5GLC

13 mars 2015, 'Thomas S. Knutsen' la3pna@gmail.com [emrfd] skrev:

 

If your crystals are at 36.86MHz, there is NO WAY they can be fundamental. The physics don't allow making crystals this high reliable. 
There is a lot of datasheets who claim that they make fundamental crystals to several hundred MHz, but measurements will show them to be overtone crystals.  

I bet your crystals will be at 12.2MHz. A good choise of a IF in a HF transceiver. Measurements show that filters made with overtone crystals have less unwanted resonances and better stopband attenuation. A fine place to get out those old 27MHz crystals and put them to use!. 

If you don't have it, get EMRFD, read the chapter on transceivers (ch. 6) and do the receiver preformance calculations. 

73 de Thomas LA3PNA.

2015-03-13 13:42 GMT+01:00 sbdick@optonline.net [emrfd] <emrfd@yahoogroups.com>:
 

Hello all. This is my first post. Allow me to introduce myself. I'm a retired electrical engineering manager who superved a digital design group. I have always had a dream of building a good performing HF communications receiver with the modern components of the day and, now that I gave the time, am proceeding with it.  I have a few questions for you, primarily for Ashar Farhan.  I am inspired by some ideas in the minima transceiver (see 

http://www.phonestack.com/farhan/minima.html) and also by the HF7070 receiver published in QEX. See:

http://www.arrl.org/files/file/QEX_Next_Issue/Jul-Aug_2013/QEX_7_13_Horrabin.pdf  I'm shooting for a balance between simplicity and good performance.  Now here's my questions:


1. A saw a mixer circuit in the VU2ESE folder for a KISS mixer/SI570/divide by 2 flop flop. Has that circuit been built and tested?  I like its simplicity


2.  I plan to use a homemade 6 pole crystal filter of the "Constant equiripple or QER GeUUR variety". I found some fundamental crystals at 29 cents each which are fundamental crystals at 36.86 mhz. I thought this might be useful for continuius general coverage but may be too lossy compared to 20 MHz crystals?  The IF would be a little too close to the upper RF frequency, making front end filtering difficult.


My head is spinning from looking at different mixewr termination techniques for best performance. (Hi dynamic range RF amps+ attenuators, diplexers, terminating the RF port at all frequencies as well as the IF port, hybrid connected dual crystal filters, etc).  I may be overthinking all the possibilities.  My current approach is to use a KISS mixer a la the Minima driving the 8 pole QER filter driving a PHA-11 amplifier chip with an attenuator pad.  The second mixer would also use a KISS mixer going down to 15 KHz, feeding a 24 bit audio A/D chip feeding a DSP IF.  For the 2 oscillators, I would use a silicon labs SI5338 for multiple outputs and low phase noise. I'd use divide by 2 flipflops on the oscillators for accurate 50% duty cycle,


Any thoughts or comments greatly appreciated.  Thank you.


"Digital Steve", K1RF





--

 Please  avoid sending  me  Word  or  PowerPoint  attachments.
 See  <http://www.gnu.org/philosophy/no-word-attachments.html>
PDF is an better alternative and there are always LaTeX!
10817 2015-03-13 10:09:32 Lasse Moell Re: Mizer front end question
OK... I found a good source describing this...
http://www.ctsvalpey.com/Collateral/Documents/English-US/Novel%20High%20Frequency%20Crystal%20Oscillator.pdf

Fundamental xtals up to 200 MHz !!!
Apparently using plano-plano crystal (flat i.e. normal) it is not economical to go above 50 MHz fundamental.... not a physical limit though.

/Lasse SM5GLC

13 mars 2015, 'Lasse Moell' lasse.moell@webdog.se [emrfd] skrev:

 

Would not be so sure about fundamental frequencies... I have been reading about crystals where the center has been etched a lot thinnner that the outer part which acts as a frame. Hence resontates at higher frequencies but not as fragile as a "normal" crystal. There has been OT crystals at ~200 MHz and not 9th but more like 3d or 5th....

I am sure other who knows this stuff will tell us all :D

/Lasse SM5GLC

13 mars 2015, 'Thomas S. Knutsen' la3pna@gmail.com [emrfd] skrev:

 

If your crystals are at 36.86MHz, there is NO WAY they can be fundamental. The physics don't allow making crystals this high reliable. 
There is a lot of datasheets who claim that they make fundamental crystals to several hundred MHz, but measurements will show them to be overtone crystals.  

I bet your crystals will be at 12.2MHz. A good choise of a IF in a HF transceiver. Measurements show that filters made with overtone crystals have less unwanted resonances and better stopband attenuation. A fine place to get out those old 27MHz crystals and put them to use!. 

If you don't have it, get EMRFD, read the chapter on transceivers (ch. 6) and do the receiver preformance calculations. 

73 de Thomas LA3PNA.

2015-03-13 13:42 GMT+01:00 sbdick@optonline.net [emrfd] <emrfd@yahoogroups.com>:
 

Hello all. This is my first post. Allow me to introduce myself. I'm a retired electrical engineering manager who superved a digital design group. I have always had a dream of building a good performing HF communications receiver with the modern components of the day and, now that I gave the time, am proceeding with it.  I have a few questions for you, primarily for Ashar Farhan.  I am inspired by some ideas in the minima transceiver (see 

http://www.phonestack.com/farhan/minima.html) and also by the HF7070 receiver published in QEX. See:

http://www.arrl.org/files/file/QEX_Next_Issue/Jul-Aug_2013/QEX_7_13_Horrabin.pdf  I'm shooting for a balance between simplicity and good performance.  Now here's my questions:


1. A saw a mixer circuit in the VU2ESE folder for a KISS mixer/SI570/divide by 2 flop flop. Has that circuit been built and tested?  I like its simplicity


2.  I plan to use a homemade 6 pole crystal filter of the "Constant equiripple or QER GeUUR variety". I found some fundamental crystals at 29 cents each which are fundamental crystals at 36.86 mhz. I thought this might be useful for continuius general coverage but may be too lossy compared to 20 MHz crystals?  The IF would be a little too close to the upper RF frequency, making front end filtering difficult.


My head is spinning from looking at different mixewr termination techniques for best performance. (Hi dynamic range RF amps+ attenuators, diplexers, terminating the RF port at all frequencies as well as the IF port, hybrid connected dual crystal filters, etc).  I may be overthinking all the possibilities.  My current approach is to use a KISS mixer a la the Minima driving the 8 pole QER filter driving a PHA-11 amplifier chip with an attenuator pad.  The second mixer would also use a KISS mixer going down to 15 KHz, feeding a 24 bit audio A/D chip feeding a DSP IF.  For the 2 oscillators, I would use a silicon labs SI5338 for multiple outputs and low phase noise. I'd use divide by 2 flipflops on the oscillators for accurate 50% duty cycle,


Any thoughts or comments greatly appreciated.  Thank you.


"Digital Steve", K1RF





--

 Please  avoid sending  me  Word  or  PowerPoint  attachments.
 See  <http://www.gnu.org/philosophy/no-word-attachments.html>
PDF is an better alternative and there are always LaTeX!
10818 2015-03-13 10:10:46 Ashhar Farhan Re: Mizer front end question
steve,

first of, there is a separate minima list that you can subscribe from minima's page. you may want to join that if you want to discuss specifics of minima. this is just so we keep emrfd focused on experiments of more generic variety.

1. yes, i did build the the KISS mixer with the flip-flop to provide the push-pull drive. i tried using the FST3157, the FSA3253, J310s, 2N7000s and even the 2N3904. Then i realized that I didn't have the proper testing facility to evaluate these. So I have spent the better part of the last 6 months building an arduino+Si570 based version of the W7ZOI spectrum analyzer. Once I finished it, I didn't go back and evaluate these variation more carefully. I will do it soon enough.

However, I must admit that the flip-flop's performance was below par. The best way to get a uniform, jitter-free push-pull is a simple voltage balun driving the two fet gates. The losses were the least with FSA3157. The best performance though came from hand-matched J310s. Joe Rocci, Steve and others have done extensive testing of these. Their mails are on the minima list and a web search will reveal their details and interesting experiments that involve JFETs, Rigol 815 SAs and cats.

2. The Minima based on the KISS mixer provides a pretty respectable IIP3 between 25dbm and 30dbm (depending upon the version of KISS mixer you use). The losses are anywhere from 13db to under 7db. Add the losses of the crystal filter (3db) and the noise figure of the IF amp (6db) and you have a noise figure of 16db : entirely usable HF. However, you might want to look at the W7ZOI's cascode IF amplifier (just one stage) for the IF amp. This can have a very low noise figure that will impact the overall noise figure of the Minima. You could realistically shoot for under 12db noise figure with the IF amp.
I don't have much experience with such high frequency fundamental crystals. We'd all love to hear about them from you. It would be simple matter to evaluate their motional parameters. At these higher frequencies, even the parallel capacitance of the crystal becomes important. You will want to include that in your design parameters.


The minima's first gain stage is the IF amplifier. It establishes sufficient gain to set the noise figure of the receiver. However, the gain shouldn't be so high that the IF amp contributes broadband noise of it's own to the detector. As Minima has no IF gain control, about 15db of gain should be sufficient given these trade-offs. Second, as the IF gain is after the main selectivity, IIP3 or IIP2 figures don't matter as much as they would if the gain stage was before the filter. 

As the audio pre-amp and the amp are where most of the receiver's gain is, it is imperative that we have a low-distortion line up. For the audio pre-amp, I'd suggest using lower noise transistors like the BC147 (2db nf) as opposed to the 2N3904 (6db nf). The audio power amplifier too needs to complement the low-distortion of the rest of the receiver. I have recently discovered that the TDA2822 and TDA2005 are excellent devices that can be used with lower gain to provide very low distortion audio output if the power levels are kept low (headphones).

For a linrad like back-end, you will need to open up the crystal filter a bit. I prefer about 5 khz bandwidth when hunting DX on CW, but that is a personal preference. 

- f

10819 2015-03-13 10:11:07 kb1gmx Re: Mizer front end question
Hi,

I've built filters using available 36.000 mhz fundamental crystals for a radio.
Right now 40mhz seems the upper frequency for fundamental rocks but above 
25mhz cheap computer clock crystals are less common and fewer choices 
in frequency.  

Those I know exist are 30, 32, 36 and 40mhz from using them.
Likely many others.

Allison
10820 2015-03-13 10:13:24 Ashhar Farhan Re: Mizer front end question
can someone please 'N2PK' these 36 mhz crystals? I don't have any 36 mhz fundamental crystals (i got a bunch of 40 mhz crystals that turned out to be 3rd overtone). 

- f

10821 2015-03-13 10:47:08 Dana Myers Re: Mizer front end question
10822 2015-03-13 10:59:00 Nick Kennedy Re: Mizer front end question
I don't know ... I tested a 28060 crystal (HC49) the other day and am pretty sure it's a fundamental.  Not too far from 36 MHz.

73-

Nick, WA5BDU

10823 2015-03-13 11:52:20 billw77aaz Re: Mizer front end question
> can someone please 'N2PK' these 36 mhz crystals? I don't have any 36 mhz
> fundamental crystals (i got a bunch of 40 mhz crystals that turned out to
> be 3rd overtone).
>
> - f

To create an optimal filter measurements should be done. But as the saying
goes "one experiment is worth....". If the crystal is put into a simple
one-transistor colpitts crystal oscillator and oscillates at 12 MHz, then
you will know whether it is an overtone crystal or fundamental.

The QER topology simplifies creation of very nice SSB-width filters since
all xtals and all coupling capacitors are identical.

W7AAZ
10825 2015-03-13 14:24:29 k1rf_digital_stev... Re: Mizer front end question
First, Thank you all for your very valuable comments.  In answer to some of your questions:

1. I did order the EMRFD book which just came via fedex today.  I am reading it voraciously.

2. I'm confident the crystals I have are fundamental. The part number is HC49S-36.864M-LF
manufactured by MEF quartz crystal. The "F" at the end indicates fundamental. I can get about a 15 KHz bandwidth with 50 ohm impedances  based on quick calculations with Dishal software. I'll get the narrow bandwidths in DSP. The wide bandwidth provides good IMD performance.

3. have built several softrocks including the ensemble II, the RxTx, and the lite II as well as the Genesis G11 so I am somewhat familiar with direct conversion receivers.  I want to use a DSP IF so I can get better image rejection than direct conversion QSDs with their small amplitude and phase errors. By sampling at a 15 KHz IF, I believe I can achieve that, like in the HF-7070.

4. I am subscribed to the minima list.

5. I planned to use FST3157s with the squarer circuit shown in the "Mixer Musings and the KISS Mixer" article by Chris Trask.

6.  The comment on the use of the SI5338 for quadrature output signals looks interesting.  I will look more closely at this.

7. I want to drive a speaker with good volume.  I have breadboarded an automotive amplifier, the STM TDA7396 which can readily drive a 4 ohm speaker with 8 watts at low distortion.  Any amp stage should be capable of bridge tied load capability (which provides 4X the power output for a given power supply voltage compared to single ended) and not need a series R-C across its output to compensate for unknown speaker impedance/electrical resonances.  As a sidenote, see my presentation on designing speakers for amateur radio at:    http://www.gnarc.net/wp-content/uploads/2014/03/k1rf_loudspeaker_design.pdf

I'm real excited about building and debugging verious sections of this receiver project.  Thanks for all your insights.

Regards, "Digital Steve", K1RF.  p.s. I apologize for calling a mixer a Mizer in the title. Fat finger syndrome.
10826 2015-03-13 17:04:57 Russell Shaw Re: Mizer front end question
10827 2015-03-13 17:37:46 billw77aaz Re: Mizer front end question
Sounds like you know what you're doing.

The FST3125 was the first "switch" chip used in mixers. There are now a
zillion different switches. Martein Bakker, PA3AKE, did extensive
performance testing on different switches. If you haven't seen it:

www.xs4all.nl/~martein/pa3ake/hmode/

Converting from 36.864 MHz to 15 KHz will require a filter with pretty
good skirt selectivity to prevent signals +30 KHz (or -30 KHz, depending
on which side the oscillator is on) from also mixing into the DSP IF. An
image reject second mixer would be neat, and reduce the requirement on the
filter. I keep thinking about it but haven't tried to build one. Yet.


W7AAZ
10828 2015-03-13 17:52:39 Gene W5DOR Re: Mizer front end question

You might want to take a look at the work of Jason Milldrum, NT7S.  He's been working with the Si5351 for some time and has a transceiver built based on this DDS.  go to nt7s.com

 

Gene, W5DOR

gene@w5dor.com

www.w5dor.com

 

 

10829 2015-03-13 21:58:43 iq_rx Re: Mizer front end question
The Drake R8 family uses a low VHF first IF with a relatively wide and simple crystal filter, then an image-reject second mixer to the low frequency IF.  All analog in the signal path, but quite elegant, and the combination of crystal filter and image-reject mixer provides nice performance.  It's an old, old technique, described in the early 1950s where the usual combination was pretty good IF selectivity followed by image-reject product detector.

I've done it a few times now when I wanted a really flat passband and graceful phase response, as it relieves the crystal filter of the burden of providing steep skirts while passing the signals you want.

Some good ideas and some depth of understanding being kicked around here.

Best Regards,

Rick KK7B
10830 2015-03-13 22:36:23 nothdurftm Re: Mizer front end question
For anyone interested there are some 38MHz crystals on ebay that are supposed to be fundamental crystals.

RALTRON AS-38.000-18-F-SMD-T 38MHz Microprocessor CRYSTAL Low Profile 10/PKG

 

K5ESS
10833 2015-03-14 09:43:19 k1rf_digital_stev... Re: Mixer front end question
Thanks for the idea. I might still use an additional mixer stage to get down to 455KHz. I could then use a cheap 6 KHz bandwidth IF filter (A Murata CFWLA455KHFA-B0, Mouser P/N 81-CFWLA455KHFA-B0) for a ridiculous ly low cost of $3.59. This actually has pretty decent skirts but is fairly wide (good for hi-fi A.M reception hihi). Then I would be able to use a single chip VGA amplifier, the Analog Devices, the AD8338. Also see Analog Devices app note AN-1192. This low current chip has a voltage controlled gain range of 0 dB to 80 dB with gain controlled in dB, can be used up to 18 MHz, Low noise: 4.5 nV/√Hz at 80 dB gain, and even has a built-in RMS detector for AGC but I wouldn't use it for AGC unless it were directly DSP controlled.  Filter bandwidth is too wide at this point.  Then I would downconvert to 15 KHz.  This if amp would have the advantage of providing most of the receiver gain and have less critical low gain audio stages with respect to low noise later in the chain. The audio stages would be used primarily for filtering, not for high gain.

Regards, "Digital Steve", K1RF
10834 2015-03-14 09:43:44 k1rf_digital_stev... Re: Mixer front end question
Hi Gene.  I am familiar with Jason Mildrum, NT7S's work and have been reading his blog.  Also the John Fisher in the AQRP Yahoo group that has developed his own software for it.  I purchased an ADAFRUIT signal generator board which I have yet to hook up to an Arduino.  I was also thinking about a function that the SI5351A contains - it is capability for spread spectrum. This is normally used for EMI reduction, but I was thinking about using that capability as a quasi narrowband noise generator for testing my crystal filter in conjunction with a softrock for looking at its spectrum.  The spread spectrum capability can provide +/- .1 to +/- 1.5 percent of operating frequency using the "center spread" mode which has a nominal 31.5 KHz spread spectrum modulation rate.  I believe this could be used as another nifty feature of that inexpensive chip.

Regards, "Digital Steve", K1RF
10835 2015-03-14 09:44:28 k1rf_digital_stev... Re: Mizer front end question
These are the crystals I purchased: 10 x 36.8640MHz 36.8640 MHz Crystal HC-49S Low Profile - ZTZSHOP -

 

not bad for 29.9 cents a crystal. I will shortly have 30 of them to play with.

"Digital Steve", K1RF
10836 2015-03-14 10:01:33 Gene W5DOR Re: Mixer front end question

Thanks for the heads up on the AQRP group.  I just joined.  I have a couple of ADAFRUIT Si5351 breakout boards as well as two of Jason's board kits.  My goal is to use all 3 outputs in a transceiver - LO,  BFO and transmit osc that will track the LO or not.

Your ideas about the spread spectrum capability is something that I hadn't thought of.  Thanks for your response.

 

Thanks & 73,

Gene W5DOR

gene@w5dor.com

www.w5dor.com

 

 

10837 2015-03-14 10:11:37 Dana Myers Re: Mixer front end question
10838 2015-03-14 10:26:04 Gene W5DOR Re: Mixer front end question

Thanks for the heads up on the cross-talk.   It would sure be better if the little device had separate Vdd pins for each output.  I could use 3 devices and add some gates to control the I2C bus.  or,  maybe use the larger device that has separate Vdd pins. 

 

Gene, W5DOR

gene@w5dor.com

www.w5dor.com

 

 

10839 2015-03-14 10:53:06 kb1gmx Re: Mizer front end question
Ashar,

They vary greatly with vendor and lot.  I did mine by hand (no N2PK handy) back then.
Filters at that frequency are generally one off unless you have a very large lot of crystals
to select from.  The biggest variance is frequency at .001% and 36mhz and 50 crystals
you might get 5 groups of 7 crystals within 100hz and each group 1khz or more different.
Also variance in ESR/Q will be wide depending on the capability and accuracy of measurement plus fixture effects.  Things you might not notice at 8mhz are significantly
greater at 36mhz.  And don't forget temperature.

That was my experience from building a 36mhz "roofing filter" actually a matched 
pair of them.

Allison
10840 2015-03-14 11:15:03 kb1gmx Re: Mizer front end question
Hi all, 

Funny not familiar with the Drake R8..

Yet I've been playing on paper with the idea of a fixed IF at VHF using analog (R2 series)
detection coupled with a image reject mixer to up convert from HF.  This was triggerd by a paper by Henderson, Cook of Watkins-Johnson.  I'll post to the KB1GMX folder.

I tend to prefer fewer conversions where possible.  One of the trouble spots in late multiple conversion radios is lack of shielding to keep the lower frequency LOs from being signals at the antenna or simply a random input to other stages.  Its part of they reason why the legendary receivers or years past used lots of metal to box everything up.  Conversations between stages needs to be by specified paths and only the signals of interest. 

Least that's the logic for a HF conversion to VHF IF then direct to baseband IQ as then the only IF/LO is above the input frequency.  If I use a image reject mixer to keep the LO+ image down (complete suppression is not required) then I should lower the spurious population and kill off a lot of possible birdies for a wide tuning radio.


Allison

10842 2015-03-14 19:30:38 vasilyivanenko Re: Mizer front end question
OK - I ordered the same and will sweep: to look a spurs, Qul and ESR
Best!
Todd
10851 2015-03-15 18:53:18 k1rf_digital_stev... Re: Mizer front end question
No joy on the crystals.  I received them but they were actually 3.6864 Mhz, not 36.864 MHz from that ebay seller.  He did give me a requested refund. My apologies to anyone who purchased these - you should request a refund. I then did further looking and found an inexpensive 40MHz fundamental crystal sold by digikey.com.  They are ECS incorporated part number ECS-400-20-46X, Digikey part number XC1769-ND. See ECS-400-20-46X ECS Inc
10860 2015-03-16 13:36:59 iam74@rocketmail.... Re: Mizer front end question
With regard to crystals, there is a seller on ebay who has some good things, AG COMPONENTS INC.

What he has is variable as he appears to be a liquidator, so one has to keep checking periodically. But I happened upon 200 pcs. of KDS 20.480 MHz xtals for ~US$9.00.

these are regular crystals from a reputable maker. I haven't yet investigated them much, but they seem to be excellent. I expect to find quite a few 10-xtal batches that are well-matched.

YRMV.

john
AD5YE