The ripple problem is not resolved, but is likely to be in concert with reducing power dissipation in adjacent components. The SAW filters need to be kept warm (-35 C) during eclipses so the second mixer needs to be replaced with something that dissipates less power. This will raise the input impedance and make the matching circuitry less critical.

73,

John KD6OZH ----- Original Message ----- From: Jim Sanford To: [email protected] Cc: [email protected] ; Bdale Garbee ; Bill Ress ; Dave Black (Home) ; Dave Black (Work) ; Dave hartzell ; David Smith ; Dick Jannson ; Don Ferguson ; John B. Stephensen ; Juan.Rivera (Work) ; Lyle Johnson ; Phil Karn ; Robert Davis KF4KSS ; Robert McGwier ; [email protected] Com ; Stephen Moraco ; 'Steven Bible' ; 'William A Schmitt' Sent: Monday, June 11, 2007 00:44 UTC Subject: Re: 70 cm Receiver Spur Source Found

Juan: Can you "sniff" the various lines from CAN-Do! to get an idea of which is the source?

Also, my memory is that we still have an open issue with the passband ripple -- or did that get resolved and I missed it?

Again, thanks for all you're doing.

73, Jim [email protected]

Juan Rivera wrote:

All,

The source of the spurs in the 70 cm prototype receiver is the CAN-Do module.

This is the receiver IF output as designed

(powered from pins 39 and 40 of the CAN-Do Module)

This is the receiver running off of a separate power source

(bypassing the CAN-Do Module +13 volt output on pins 39 and 40)

This is the same configuration as above, but with the CAN-Do module shut down

From the above it appears that:

<!--[if !supportLists]-->1. <!--[endif]-->The CAN-Do module is the source of all the spurs that I see

<!--[if !supportLists]-->2. <!--[endif]-->None of the spurs are being generated by the receiver's internal power supplies

<!--[if !supportLists]-->3. <!--[endif]-->Some, but not all, of the noise is entering the receiver via the switched +13 volt CAN-Do output on pins 39 and 40

<!--[if !supportLists]-->4. <!--[endif]-->The remainder of the noise could be entering the receiver on other CAN-Do pins, or being radiated to receiver components or traces in close proximity

See my log for all the intermediate steps I took to reach these conclusions.

Remaining RF issues:

<!--[if !supportLists]-->1. <!--[endif]-->Phase noise

<!--[if !supportLists]-->2. <!--[endif]-->Microphonics

73,

Juan

WA6HTP

It meets the requirements from more than a year ago that are posted on the U-band receiver page. Matt and Phil can comment on the text-mode uplink requirements.

73,

John KD6OZH ----- Original Message ----- From: Jim Sanford To: [email protected] Cc: [email protected] ; Bdale Garbee ; Bill Ress ; Dave Black (Home) ; Dave Black (Work) ; Dave hartzell ; David Smith ; Dick Jannson ; Don Ferguson ; John B. Stephensen ; Juan.Rivera (Work) ; Lyle Johnson ; Phil Karn ; Robert Davis KF4KSS ; Robert McGwier ; [email protected] Com ; Stephen Moraco ; 'Steven Bible' ; 'William A Schmitt' Sent: Monday, June 11, 2007 00:53 UTC Subject: Re: 70 cm Receiver Spur Source Found

All: Slightly off topic . . . .

Somewhere in this discussion, Juan has asked basically, "what is acceptable phase noise at what offset?" This is similar to a question I asked Bob McGwier some months ago, as I contemplate various options for my 6m to 10 GHz terrestrial station. The question arises because so many texts have different opinions and the manufacturers each choose a different offset for their specs, carefully avoiding evaluation of the phase noise of their device on whatever communications mode.

I think I remember Bob's answer as -110dbc at 10 KHz offset.

If I remember Bob's answer correctly (Which I may NOT, Bob, please correct me) then the picture Juan shows, which is the composite of his signal generator, his SDRIQ, and the U-band downconverter doesn't look too bad -- close, but maybe not close enough, but maybe so. John, Matt, & Phil: What is your evaluation of this phase noise on our intended uses of the receiver?

Thanks & 73, Jim [email protected]

Juan Rivera wrote:

All,

The source of the spurs in the 70 cm prototype receiver is the CAN-Do module.

This is the receiver IF output as designed

(powered from pins 39 and 40 of the CAN-Do Module)

Bill: -110 dbc/hz at 10khz offset is what I REMEMBER in my response to the question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?"

My memory may very well be WRONG, which was the point of asking my question.

I agree with you, looks like -90dbC at 10 KHz, but given what Juan's already said about his sources, and the uncertainties of the SDR-IQ, that may be VERY good!

Do you have a different answer to my original question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?" If so, I'd like to engage in that conversation, both for Eagle and for my terrestrial microwave station.

I agree with your assessment of Juan's work -- ABSOLUTELY EXEMPLARY!!!!!!

Thanks & 73, Jim [email protected]

Bill Ress wrote:

Jim,

Where did the -110 dBC/Hz at 10 KHz come from? And at what frequency? That's pretty aggressive using straight forward VCO's and PLLs. From Juan's data with his snappy SDRIQ, it looks like that the 70 cm is doing about -90 dBc/Hz at10 KHz. I thought that's what I read from his early HP8566B dtata.

My straight forward S2 RX downconverter PLL LO does -100 dBc/Hz at 10 KHz.

We'd better get a good read on this since, if real, it forces more sophisticated LO designs.

Great work Juan!!

Regards...Bill - N6GHz

Jim Sanford wrote:

...

I think I remember Bob's answer as -110dbc at 10 KHz offset.

Thanks & 73, Jim [email protected]

On Sun, 2007-06-10 at 20:04 -0700, Juan Rivera wrote:

To my surprise the module jumps briefly to about 57 kHz at regular intervals of about 3 seconds.

Yes, that's the watchdog timer firing and causing a software reset. I have no doubt that the reset cycle causes a short-term variation in the current consumption profile on the board itself. If the frequency is power consumption dependent, note that it's likely to move around a bit during normal operation... might want at some point to look at conditions with the CAN-Do! running at a "normal" 20ms update rate.

My next plan is to probe all 40 pins by using a 10x scope probe. I’ll start another log called CAN-Do Noise and start posting there as I start on this task.

As Lyle pointed out in an off-list email, the power supply on the CAN-Do! widget was designed to be efficient, not necessarily to be quiet. Switchers always have noise issues... we've flown a lot of them, and this isn't the first time some filtering or shielding has been required to keep everything happy. You're doing the right things to characterize what's going on, and we'll all help if/as needed.

We (the CAN-Do folk) are following your progress with great interest. Augmenting the CAN-Do! documentation with some summary of your findings and mitigation approach to aid future users is something we'll look forward to doing.

Bdale

Not to mention that ALL power supplies to the oscillators were specified to have adequate filtering and noise immunity. We need to revisit that. There is NO power supply coming from the spacecraft bus where you may assume it is nice, flat, and quiet. Assume it is ugly, raw, and must be conditioned.

Bob

Bdale Garbee wrote:

On Sun, 2007-06-10 at 20:04 -0700, Juan Rivera wrote:

...

To my surprise the module jumps briefly to about 57 kHz at regular intervals of about 3 seconds.

Yes, that's the watchdog timer firing and causing a software reset. I have no doubt that the reset cycle causes a short-term variation in the current consumption profile on the board itself. If the frequency is power consumption dependent, note that it's likely to move around a bit during normal operation... might want at some point to look at conditions with the CAN-Do! running at a "normal" 20ms update rate.

...

My next plan is to probe all 40 pins by using a 10x scope probe. I’ll start another log called CAN-Do Noise and start posting there as I start on this task.

As Lyle pointed out in an off-list email, the power supply on the CAN-Do! widget was designed to be efficient, not necessarily to be quiet. Switchers always have noise issues... we've flown a lot of them, and this isn't the first time some filtering or shielding has been required to keep everything happy. You're doing the right things to characterize what's going on, and we'll all help if/as needed.

We (the CAN-Do folk) are following your progress with great interest. Augmenting the CAN-Do! documentation with some summary of your findings and mitigation approach to aid future users is something we'll look forward to doing.

Bdale

Bdale,

Please refresh me on the function of the watchdog timer. In the state I was running, while observing the 3-second jumps, the receiver was shut down but the CAN-Do module and dongle were both powered up (pins 39 and 40 not connected to the receiver.) I thought the module only reset if it detected no heart beat from the dongle.

I think this evening I will fire up the receiver to recreate the noise I've been seeing and then start playing with some brass sheet in front of the inductor as a shield. I tried probing all the pins with a 10x scope probe connected to the SDR-IQ but I didn't see anything. I'll repeat that test with a 1x probe. Based on last night I'm currently thinking that the noise is radiated from that inductor rather than conducted via the 40-pin connector. There are receiver components very close to that inductor.

73,

Juan WA6HTP

On 6/11/07, Bdale Garbee [email protected] wrote:

On Sun, 2007-06-10 at 20:04 -0700, Juan Rivera wrote:

...

To my surprise the module jumps briefly to about 57 kHz at regular intervals of about 3 seconds.

Yes, that's the watchdog timer firing and causing a software reset. I have no doubt that the reset cycle causes a short-term variation in the current consumption profile on the board itself. If the frequency is power consumption dependent, note that it's likely to move around a bit during normal operation... might want at some point to look at conditions with the CAN-Do! running at a "normal" 20ms update rate.

...

My next plan is to probe all 40 pins by using a 10x scope probe. I'll start another log called CAN-Do Noise and start posting there as I start on this task.

As Lyle pointed out in an off-list email, the power supply on the CAN-Do! widget was designed to be efficient, not necessarily to be quiet. Switchers always have noise issues... we've flown a lot of them, and this isn't the first time some filtering or shielding has been required to keep everything happy. You're doing the right things to characterize what's going on, and we'll all help if/as needed.

We (the CAN-Do folk) are following your progress with great interest. Augmenting the CAN-Do! documentation with some summary of your findings and mitigation approach to aid future users is something we'll look forward to doing.

Bdale

Juan,

A further clarification after re-reading your email.

The dongle is the Lawicel CAN232 device. It is simply an RS232 to CAN adapter and is not capable of generating any CAN traffic on its own.

The CDNC software when configured to do so will tell the CAN232 what packet content to send over the CAN bus and when.

So, the way to think about this is there is no traffic/hearbeat unless the CDNC software is generating it.

Set the heartbeat rate and then enable the hearbeat in CDNC and you should have traffic preventing Watchdog resets.

This help clarifiy?

Regards, Stephen --

Juan: I have seen, in my professional life, a similar situation where an inductor was radiating noise with unacceptable impact. Replacing the inductor with a shielded version was more effective than any overall shield, tho those did work.

Good luck, and thanks again for all your hard work.

73, Jim [email protected]

Juan Rivera wrote:

Bdale,

Please refresh me on the function of the watchdog timer. In the state I was running, while observing the 3-second jumps, the receiver was shut down but the CAN-Do module and dongle were both powered up (pins 39 and 40 not connected to the receiver.) I thought the module only reset if it detected no heart beat from the dongle.

I think this evening I will fire up the receiver to recreate the noise I've been seeing and then start playing with some brass sheet in front of the inductor as a shield. I tried probing all the pins with a 10x scope probe connected to the SDR-IQ but I didn't see anything. I'll repeat that test with a 1x probe. Based on last night I'm currently thinking that the noise is radiated from that inductor rather than conducted via the 40-pin connector. There are receiver components very close to that inductor.

73,

Juan WA6HTP

On 6/11/07, *Bdale Garbee* <[email protected] mailto:[email protected]> wrote:

On Sun, 2007-06-10 at 20:04 -0700, Juan Rivera wrote:
> To my surprise the module jumps briefly to about 57 kHz at regular
> intervals of about 3 seconds.

Yes, that's the watchdog timer firing and causing a software reset.  I
have no doubt that the reset cycle causes a short-term variation
in the
current consumption profile on the board itself.  If the frequency is
power consumption dependent, note that it's likely to move around
a bit
during normal operation... might want at some point to look at
conditions with the CAN-Do! running at a "normal" 20ms update rate.

> My next plan is to probe all 40 pins by using a 10x scope
probe.  I'll
> start another log called CAN-Do Noise and start posting there as I
> start on this task.

As Lyle pointed out in an off-list email, the power supply on the
CAN-Do! widget was designed to be efficient, not necessarily to be
quiet.  Switchers always have noise issues... we've flown a lot of
them,
and this isn't the first time some filtering or shielding has been
required to keep everything happy.   You're doing the right things to
characterize what's going on, and we'll all help if/as needed.

We (the CAN-Do folk) are following your progress with great interest.
Augmenting the CAN-Do! documentation with some summary of your
findings
and mitigation approach to aid future users is something we'll look
forward to doing.

Bdale

On Mon, 2007-06-11 at 18:54 -0700, Juan Rivera wrote:

It might be nice to know what steps have been taken in the past to reduce CAN-Do noise. Stephen?

We've had very little feedback from the AMSAT-DL users of CAN-Do!, and the only P3E spacecraft module that I can confirm first-hand they've actually interfaced with a CAN-Do! widget is a sensor electronics unit that would not have the same sensitivity to this issue as a receiver.

I suspect, therefore, that you're blazing a new trail. ;-)

Bdale

Hi Guys,

There are issues with Tantalums. From my perspective (and I could be very WRONG!!!!) Lyle, and Chuck Green are AMSAT's corporate knowledge on space-suitable components. We should make no changes without consulting them, and possibly Jan King as well. Lou McFadin may have some useful insight, also.

Don't look for consensus on this topic. Like most subjects, you can find "experts" to support any position you want to take.

We've used a lot of Tantalums with good success. My choice is to use Tantalum rather than Electrolytic whenever possible. But I lay out PCBs and populate them, hence my bias. I don't normally have to make the circuit work.

The position taken by Karl Meinzer is that you should never use Tantalums. The argument I have heard to support this is that the failure mode of a Tantalum is a short. Of course, depending on the circuit, if the capacitor fails, it might not make much difference what the failure mode is. But for many circuits, if the capacitor fails, the circuit will probably keep working if the capacitor does not short. I have read that this is no longer a problem for modern parts (newer than about 10 years) because it is no longer the failure mode. I do have one recent experience with this when I put a part in backward which resulted in the predictable explosion. The failure mode... shorted.

Note that the filter capacitors on the CAN-Do! are not Tantalum (except one). This is because of Karl's rule. He did allow the one when we severally de-rated the voltage rating.

Electrolytic capacitors have their own problems. The biggest is their tendency to dry out, especially in a vacuum. We address this by sealing the open end of the can. Yes, this creates the potential for an even bigger bang, but we haven't seen this yet. Epoxy has been used. I also use the conformal coating for this purpose as appropriate. You are not likely to use conformal coating on an RF circuit because of the bad things it's dialectic constant will likely do to your carefully tuned circuits.

My advice is to choose the technology that best suits your needs. They both work and we have a lot of success with both of them when we take precautions as described above.

Chuck

The high-capacitance ceramic capacitors are very useful for switching regulators and for general bypassing. They're actually less expensive than tantalum electrolytics in most cases. Consequently, there are no tantalum capacitors in the 70 cm receiver.

NP0 capacitors are probably OK, but a 0.01 uF capacitor in an 0603 size appears to be much more microphonic than the larger leaded version. I used plastic film capacitors in the PLL integrators but should have used them everywhere in the feedback loops. Unfortunately, they are going to be larger than the ceramic capacitors.

73,

John KD6OZH

----- Original Message ----- From: "Matt Ettus" [email protected] To: "Chuck Green" [email protected] Cc: [email protected] Sent: Tuesday, June 12, 2007 05:05 UTC Subject: [eagle] Re: Jim's Phase Noise Number!!

...

My advice is to choose the technology that best suits your needs. They both work and we have a lot of success with both of them when we take precautions as described above.

I don't know how they do in space, but you can now get very large ceramic caps -- like 22uF in a 1206 package.

Matt

---

Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

I've updated the requirements document on the U-band receiver web page. This version accounts for text-mode uplinks but that doesn't change the requirements significantly. I did make assumptions about the antenna pattern (the same as before) and the text-mode uplink spacing and bandwidth (these are new). I've allowed for splitting the receiver into two modules in case this is needed to solve the thermal problem. I've also changed the requirements for PAVE PAWS interference abatement by assuming that the SAW filters will be the primary mechanism used. This may allow lower DC power consumption.

All of the requirements should be verifed by those interfacing to the receiver.

73,

John KD6OZH

All the ceramic capacitors in the PLL feedback path that aren't NP0 types should be replaced with film capacitors. I'll see what size is available to replace C19, C27, C28, C50, C51 and C54.

73,

John KD6OZH ----- Original Message ----- From: Juan Rivera To: 'Jim Sanford' Cc: [email protected] Sent: Tuesday, June 12, 2007 01:54 UTC Subject: [eagle] Re: Jim's Phase Noise Number!!

Thanks Jim,

I spent some time this evening trying to dig into the CAN-Do noise. I was able to reduce it by 3 or 4 dB by temporarily shielding the module with aluminum tape. Before getting too excited we need to answer the question - is this noise going to be a problem? I think as a general suggestion we should move the RF to the far end of the enclosure. We have the space. I think it would be best to just extend the traces and not get involved trying to decouple the leads with ferrite beads and a cable - or perhaps build in some filtering into the PCB and implement with SMD components. Extend the input and output traces so we can stick with the three PCB mount SMA connectors. It might be nice to know what steps have been taken in the past to reduce CAN-Do noise. Stephen?

I have another idea that would involve a milled enclosure with multiple compartments. I can see a modular structure where the sections are assembled into a chassis and pulled together with threaded rods. The individual sections would be kept in alignment with pins. That would take care of all the RFI and mechanical flexing issues. If you're willing to consider this idea I'll follow up with details. Unless the existing chassis is made rigid and plumb it is a bit concern of mine.

Moving to microphonics. I did a quick check of available information on the Internet and found this: "Barium titanate, which is the base ceramic material for most dielectric systems, will exhibit microphonic effects." That article went on to suggest tantalum as a non-microphonic choice. Since it should be pretty quiet out there in orbit microphonics shouldn't be a problem. I did find out where the bulk of is coming from however - C27 and C28. You'll find them on schematic page 4 between U4 and U5. Almost every capacitor on the board appears to exhibit some microphonics but those two seem to be the worst.

73,

Juan

------------------------------------------------------------------------------

From: Jim Sanford [mailto:[email protected]] Sent: Monday, June 11, 2007 6:21 PM To: Juan Rivera Cc: Bdale Garbee; Bill Ress; [email protected] Subject: Re: [eagle] Re: Jim's Phase Noise Number!!

Juan: I have seen, in my professional life, a similar situation where an inductor was radiating noise with unacceptable impact. Replacing the inductor with a shielded version was more effective than any overall shield, tho those did work.

Good luck, and thanks again for all your hard work.

73, Jim [email protected]

Juan Rivera wrote:

Bdale,

Please refresh me on the function of the watchdog timer. In the state I was running, while observing the 3-second jumps, the receiver was shut down but the CAN-Do module and dongle were both powered up (pins 39 and 40 not connected to the receiver.) I thought the module only reset if it detected no heart beat from the dongle.

I think this evening I will fire up the receiver to recreate the noise I've been seeing and then start playing with some brass sheet in front of the inductor as a shield. I tried probing all the pins with a 10x scope probe connected to the SDR-IQ but I didn't see anything. I'll repeat that test with a 1x probe. Based on last night I'm currently thinking that the noise is radiated from that inductor rather than conducted via the 40-pin connector. There are receiver components very close to that inductor.

73,

Juan

WA6HTP

On 6/11/07, Bdale Garbee [email protected] wrote:

On Sun, 2007-06-10 at 20:04 -0700, Juan Rivera wrote:

To my surprise the module jumps briefly to about 57 kHz at regular intervals of about 3 seconds.

Yes, that's the watchdog timer firing and causing a software reset. I have no doubt that the reset cycle causes a short-term variation in the current consumption profile on the board itself. If the frequency is power consumption dependent, note that it's likely to move around a bit during normal operation... might want at some point to look at conditions with the CAN-Do! running at a "normal" 20ms update rate.

My next plan is to probe all 40 pins by using a 10x scope probe. I'll start another log called CAN-Do Noise and start posting there as I start on this task.

As Lyle pointed out in an off-list email, the power supply on the CAN-Do! widget was designed to be efficient, not necessarily to be quiet. Switchers always have noise issues... we've flown a lot of them, and this isn't the first time some filtering or shielding has been required to keep everything happy. You're doing the right things to characterize what's going on, and we'll all help if/as needed.

We (the CAN-Do folk) are following your progress with great interest. Augmenting the CAN-Do! documentation with some summary of your findings and mitigation approach to aid future users is something we'll look forward to doing.

Bdale

------------------------------------------------------------------------------

_______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

Juan, Why not think out of the box, Put the Can Do module outside the box, perhaps on top of the module. I think that is a more likely solution than re building the Can-do module.

Lou McFadin W5DID [email protected]

On Jun 29, 2007, at 10:25 PM, Juan Rivera wrote:

Bdale,

I'm sorry to pick on you, but you seem like a good guy to complain to...

I seem to be having a problem stimulating a discussion. After saying that I though the CAN-Do power supply needed to be completely scrapped and replaced with one running at around 1 MHz I expected to be buried in email, but I've only received one message referring to this so far, and it wasn't from anyone working on CAN-Do.

The CAN-Do module is unique in that it is going to be an integral component INSIDE of every payload, so any deficiencies it may have are going to have a large impact.

Let me restate my finding so far:

There are 4 categories of EMI and the CAN-Do module / 70 cm Receiver combination is experiencing all four. They break down into radiated and conducted emissions and susceptibility, and they are generally referred to by a two-letter designation - RE, RS, CE, and CS.

RE and RS go together - the CAN-Do module's switching power supply inductor radiates the 5 kHz switching noise out the back directly towards the receiver (RE.) The Receiver's VCO's are both very sensitive to EMI and are impacted by the CAN-Do module if they are within 4-1/2 inches of the inductor (RS.) I've had to move the CAN-Do module off of the receiver PCB and interconnect it with a ribbon cable to deal with this problem. The good news is that I am fairly confident that It can be fixed by going to a two-compartment chassis, with a bulkhead separating the CAN-Do module from the analog Receiver to provide shielding. The bad news is that I think this means we need a milled module chassis.

CE and CS also go together, and this is the real problem I see since you can't fix conducted EMI with a shielded enclosure. It requires filtering. The CAN-Do module is trashing the DC input from the power source and also feeding noisy power to the Receiver. The outgoing noise is the bigger concern because it will add to the CS problems for everything connected to the power source. In the other direction, the switched power from the CAN-Do module shows up in the IF output as 5 kHz spurs. Moving the CAN-Do module physically away from the Receiver only dealt with the RE/RS issue. I had to bypass the CAN-Do module and run clean power directly from the lab bench supply to deal with the CE/CS problem. This means that there is no current monitoring and no power control.

The 5 kHz switching frequency is bad for two reasons - it makes filtering this noise a much larger problem than it needs to be, and the impact is more severe since it is putting spurs all over the passband of the IF at 5 kHz intervals.

If you sit back and think about the impact of a dozen noisy power supplies all feeding EMI back to the common power source where they all mix together and make their way back to each payload, it starts to look nasty. All these supplies will be drifting around and beating with each other to produce sum and difference noise on the power bus. 5 kHz noise is hard enough to get rid of but what if there are difference components at a few hundred Hz? How can you design a filter when you won't know what to expect until you hook everything up and turn it on? And by then you're out of time.

I'm not making this stuff up. People I know have run into this exact problem before and the result was very bad.

A while ago Howard Long made a suggestion that I think has great potential. Here's what he had to say:

...in the original SDX PSU design I had in San Francisco last year is an SMPS using the LM2672 device. These can be fitted with an AC coupled SYNC signal to override the internal default SMPS frequency. I selected 375kHz for my unit (6MHz divided by 16) to ensure its harmonics were outside the 10.7MHz IF passband. If the external SYNC fails the internal SMPS oscillator takes over.

My Suggestions:

1. Revise the CAN-Do module to move the switching frequency up as

far as possible to move spurs out of the passband of sensitive analog circuitry, and to ease the burden on EMI filtering. 2) The power and grounds must be filtered in both directions to minimize CE back to the power source and to the payload. 3) The switching inductor should be a shielded to reduce RE inside the module chassis. 4) A power supply capable of synching to a master oscillator should be strongly considered.

This topic needs to be elevated to the top of the queue. The EMI environment surrounding the CAN-Do module impacts the design of the next revision of the 70 cm Receiver, and also directly impacts the chassis design. What do we need to do to get going on this?

73,

Juan - WA6HTP

---

Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

Hi John,

I had a chat or two with Bob Davis about chassis and I think you can bring SMA connectors out one side and the CAN-Do out the end, but nothing out the back.

Given the amount of conducted 5 kHz noise I'm seeing, do you think you can filter it out? Don't forget, I'm feeding the receiver with bypassed clean DC from the bench supply at the moment. The CAN-Do switched DC is nasty.

In your next version why not raise the switching frequency up and get it out of the passband completely? What do you think of Howard's idea of synching the switching supplies?

Juan

_____

From: John B. Stephensen [mailto:[email protected]] Sent: Friday, June 29, 2007 9:24 PM To: Louis McFadin; [email protected] Cc: Dave Black (Work); Dave Black (Home); David Smith; [email protected]; [email protected] Com; Juan.Rivera (Work) Subject: U-band receiver changes to reduce EMI and improve thermal regulation

For the next version of the 70 cm receiver, I'm thinking of an arrangement where the CAN bus exits out of the opposite side of the module from the RF and IF. Does this create any problems in the wiring harness?

The CAN-DO module would be mounted on a small PCB containing most of the power supply circuitry. This PCB would contain the switcher that generates 7 VDC for the receiver and some of the linear regulators. It would be mounted at one end of the module and be heat-sinked.

The RF circuitry would be mounted on a separate PCB at the other end of the module. This allows all RF and IF coax connectors to mount on this PCB and attach directly to the connector mounting bracket. It also eliminates the flying lead for the frequency reference input. The power dissipation would be limited so that it doesn't need to be heat-sinked. This protects the SAW filters from rapid temperature excursions and keeps them above -30 C at all times.

The two PCBs would be connected with a cable carrying DC power and the signals being monitored. Each PCB would have filtering for power and the signals being monitored by the IHU. The interconnect cable would travel through a common-mode choke using high-permiability ferrite.

If necessary for thermal reasons, the RF amplifier and first mixer would go on a third PCB with a heat sink. LO and IF would connect to the second PCB via 2 coax cables. Once the new requirements document is approved, the power dissipation can be calculated for each of the 3 PCBs and a thermal analysis would determine whether 2 or 3 PCBs are needed and how far apart they should be.

Unless more problems are found during testing of the current version of the receiver, the circuitry would be same as now, except that an MCU is added to initialize the PLLs and the second mixer and second IF amplifiers are changed to devices that dissipate less power in line with the new requirements document. Bob has expressed some interest in Peregrine PLLs that can have the frequency hard wired so they could be inserted instead of adding the MCU.

73,

John

KD6OZH

----- Original Message -----

From: Louis McFadin mailto:[email protected]

To: [email protected]

Cc: Dave mailto:[email protected] Black (Work) ; Dave Black (Home) mailto:[email protected] ; David Smith mailto:[email protected] ; [email protected] ; [email protected] Com mailto:[email protected]%20Com ; Juan.Rivera mailto:[email protected] (Work)

Sent: Saturday, June 30, 2007 02:36 UTC

Subject: [eagle] Re: CAN-Do EMI - Let's Get Going on This!

Juan,

Why not think out of the box, Put the Can Do module outside the box, perhaps on top of the module.

I think that is a more likely solution than re building the Can-do module.

Lou McFadin

W5DID

[email protected]

On Jun 29, 2007, at 10:25 PM, Juan Rivera wrote:

Bdale,

I'm sorry to pick on you, but you seem like a good guy to complain to...

I seem to be having a problem stimulating a discussion. After saying that I

though the CAN-Do power supply needed to be completely scrapped and replaced

with one running at around 1 MHz I expected to be buried in email, but I've

only received one message referring to this so far, and it wasn't from

anyone working on CAN-Do.

The CAN-Do module is unique in that it is going to be an integral component

INSIDE of every payload, so any deficiencies it may have are going to have a

large impact.

Let me restate my finding so far:

There are 4 categories of EMI and the CAN-Do module / 70 cm Receiver

combination is experiencing all four. They break down into radiated and

conducted emissions and susceptibility, and they are generally referred to

by a two-letter designation - RE, RS, CE, and CS.

RE and RS go together - the CAN-Do module's switching power supply inductor

radiates the 5 kHz switching noise out the back directly towards the

receiver (RE.) The Receiver's VCO's are both very sensitive to EMI and are

impacted by the CAN-Do module if they are within 4-1/2 inches of the

inductor (RS.) I've had to move the CAN-Do module off of the receiver PCB

and interconnect it with a ribbon cable to deal with this problem. The good

news is that I am fairly confident that It can be fixed by going to a

two-compartment chassis, with a bulkhead separating the CAN-Do module from

the analog Receiver to provide shielding. The bad news is that I think this

means we need a milled module chassis.

CE and CS also go together, and this is the real problem I see since you

can't fix conducted EMI with a shielded enclosure. It requires filtering.

The CAN-Do module is trashing the DC input from the power source and also

feeding noisy power to the Receiver. The outgoing noise is the bigger

concern because it will add to the CS problems for everything connected to

the power source. In the other direction, the switched power from the

CAN-Do module shows up in the IF output as 5 kHz spurs. Moving the CAN-Do

module physically away from the Receiver only dealt with the RE/RS issue. I

had to bypass the CAN-Do module and run clean power directly from the lab

bench supply to deal with the CE/CS problem. This means that there is no

current monitoring and no power control.

The 5 kHz switching frequency is bad for two reasons - it makes filtering

this noise a much larger problem than it needs to be, and the impact is more

severe since it is putting spurs all over the passband of the IF at 5 kHz

intervals.

If you sit back and think about the impact of a dozen noisy power supplies

all feeding EMI back to the common power source where they all mix together

and make their way back to each payload, it starts to look nasty. All these

supplies will be drifting around and beating with each other to produce sum

and difference noise on the power bus. 5 kHz noise is hard enough to get

rid of but what if there are difference components at a few hundred Hz? How

can you design a filter when you won't know what to expect until you hook

everything up and turn it on? And by then you're out of time.

I'm not making this stuff up. People I know have run into this exact

problem before and the result was very bad.

A while ago Howard Long made a suggestion that I think has great potential.

Here's what he had to say:

...in the original SDX PSU design I had in San Francisco last year is an

SMPS using the LM2672 device. These can be fitted with an AC coupled SYNC

signal to override the internal default SMPS frequency. I selected 375kHz

for my unit (6MHz divided by 16) to ensure its harmonics were outside the

10.7MHz IF passband. If the external SYNC fails the internal SMPS oscillator

takes over.

My Suggestions:

1) Revise the CAN-Do module to move the switching frequency up as far as

possible to move spurs out of the passband of sensitive analog circuitry,

and to ease the burden on EMI filtering.

2) The power and grounds must be filtered in both directions to minimize CE

back to the power source and to the payload.

3) The switching inductor should be a shielded to reduce RE inside the

module chassis.

4) A power supply capable of synching to a master oscillator should be

strongly considered.

This topic needs to be elevated to the top of the queue. The EMI

environment surrounding the CAN-Do module impacts the design of the next

revision of the 70 cm Receiver, and also directly impacts the chassis

design. What do we need to do to get going on this?

73,

Juan - WA6HTP

_______________________________________________

Via the Eagle mailing list courtesy of AMSAT-NA

[email protected]

http://amsat.org/mailman/listinfo/eagle

_____

_______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

John,

For this particular application I think that approach might work, but it forces others to make major concessions in terms of utilization of space inside the enclosure to avoid CAN-Do radiated EMI. Regardless of what happens to the CAN-Do EMI it looks like you better plan on some nasty DC power and filter the DC input accordingly.

I can see a small PCB attached to the CAN-Do 40-pin connector that contains DC filtering, the receiver's switching power supply, and lands for the signal lines that need to go back to the receiver (not all 40 conductors - just the ones we actually use.) We'd probably want to run those wires through to the other compartment via feed-thru filters in the common bulkhead using a small wiring harness, then another small harness on the other side to interface to the receiver PCB. An alternate approach might be to use a smaller ribbon cable and connectors to move the actual signals and power back to the receiver, but that adds weight and doesn't allow for filtering through the bulkhead unless we found an EMI ribbon cable bulkhead connector. I think I'd prefer the discrete wire harness approach. I think it's lighter, more reliable, and eliminates one set of connectors.

Before proceeding with another revision I'd like to see an EMI requirement spec and a practical way to generate the expected noisy DC power. This will also give the next group something to design and test to.

By the way, did you see my phase noise data? I went back and tested again using another spectrum analyzer while mine is getting calibrated. You'll notice an interesting diversion between the two. As the saying goes, "A man with a watch always knows what time it is. A man with two watches can never be sure." I'll run that again when I get the calibrated spec. a. back from the shop.

73,

Juan

_____

From: John B. Stephensen [mailto:[email protected]] Sent: Friday, June 29, 2007 10:38 PM To: [email protected]; 'Louis McFadin' Cc: 'Dave Black (Work)'; 'Dave Black (Home)'; 'David Smith'; [email protected]; '[email protected] Com'; 'Juan.Rivera (Work)' Subject: Re: U-band receiver changes to reduce EMI and improve thermal regulation

If the CAN-Do module can be placed in a shielded compartment with multi-pole filters on all wires leaving the shielded area, we can get at least 80 dB of attenuation for conducted interference and the electric field component of radiated interference. Magnetic shielding is harder as it requires iron or mu-metal so a magnetically shielded inductor should be used in the CAN-DO module.

Synchronizing the switchers to a frequency with no harmonics at 10.5-10.9 MHz would help with radiated interference, but a shielded inductor may be sufficient if it can be moved far enough from the RF circuitry -- especially the PLLS and VCOs.

If we have only two adjacent sides available for connectors, the CAN-Do module and power supply circuitry could attach to the existing connector bracket. The RF and IF connectors would then exit along one long side of the box as far from the CAN bus connector as possible.

73,

John

KD6OZH

----- Original Message -----

From: Juan mailto:[email protected] Rivera

To: 'John B. mailto:[email protected] Stephensen' ; 'Louis mailto:[email protected] McFadin'

Cc: 'Dave mailto:[email protected] Black (Work)' ; 'Dave Black (Home)' mailto:[email protected] ; 'David Smith' mailto:[email protected] ; [email protected] ; '[email protected] mailto:'[email protected]%20Com' Com' ; 'Juan.Rivera (Work)' mailto:[email protected]

Sent: Saturday, June 30, 2007 04:32 UTC

Subject: RE: U-band receiver changes to reduce EMI and improve thermal regulation

Hi John,

I had a chat or two with Bob Davis about chassis and I think you can bring SMA connectors out one side and the CAN-Do out the end, but nothing out the back.

Given the amount of conducted 5 kHz noise I'm seeing, do you think you can filter it out? Don't forget, I'm feeding the receiver with bypassed clean DC from the bench supply at the moment. The CAN-Do switched DC is nasty.

In your next version why not raise the switching frequency up and get it out of the passband completely? What do you think of Howard's idea of synching the switching supplies?

Juan

_____

From: John B. Stephensen [mailto:[email protected]] Sent: Friday, June 29, 2007 9:24 PM To: Louis McFadin; [email protected] Cc: Dave Black (Work); Dave Black (Home); David Smith; [email protected]; [email protected] Com; Juan.Rivera (Work) Subject: U-band receiver changes to reduce EMI and improve thermal regulation

For the next version of the 70 cm receiver, I'm thinking of an arrangement where the CAN bus exits out of the opposite side of the module from the RF and IF. Does this create any problems in the wiring harness?

The CAN-DO module would be mounted on a small PCB containing most of the power supply circuitry. This PCB would contain the switcher that generates 7 VDC for the receiver and some of the linear regulators. It would be mounted at one end of the module and be heat-sinked.

The RF circuitry would be mounted on a separate PCB at the other end of the module. This allows all RF and IF coax connectors to mount on this PCB and attach directly to the connector mounting bracket. It also eliminates the flying lead for the frequency reference input. The power dissipation would be limited so that it doesn't need to be heat-sinked. This protects the SAW filters from rapid temperature excursions and keeps them above -30 C at all times.

The two PCBs would be connected with a cable carrying DC power and the signals being monitored. Each PCB would have filtering for power and the signals being monitored by the IHU. The interconnect cable would travel through a common-mode choke using high-permiability ferrite.

If necessary for thermal reasons, the RF amplifier and first mixer would go on a third PCB with a heat sink. LO and IF would connect to the second PCB via 2 coax cables. Once the new requirements document is approved, the power dissipation can be calculated for each of the 3 PCBs and a thermal analysis would determine whether 2 or 3 PCBs are needed and how far apart they should be.

Unless more problems are found during testing of the current version of the receiver, the circuitry would be same as now, except that an MCU is added to initialize the PLLs and the second mixer and second IF amplifiers are changed to devices that dissipate less power in line with the new requirements document. Bob has expressed some interest in Peregrine PLLs that can have the frequency hard wired so they could be inserted instead of adding the MCU.

73,

John

KD6OZH

----- Original Message -----

From: Louis McFadin mailto:[email protected]

To: [email protected]

Cc: Dave mailto:[email protected] Black (Work) ; Dave Black (Home) mailto:[email protected] ; David Smith mailto:[email protected] ; [email protected] ; [email protected] Com mailto:[email protected]%20Com ; Juan.Rivera mailto:[email protected] (Work)

Sent: Saturday, June 30, 2007 02:36 UTC

Subject: [eagle] Re: CAN-Do EMI - Let's Get Going on This!

Juan,

Why not think out of the box, Put the Can Do module outside the box, perhaps on top of the module.

I think that is a more likely solution than re building the Can-do module.

Lou McFadin

W5DID

[email protected]

On Jun 29, 2007, at 10:25 PM, Juan Rivera wrote:

Bdale,

I'm sorry to pick on you, but you seem like a good guy to complain to...

I seem to be having a problem stimulating a discussion. After saying that I

though the CAN-Do power supply needed to be completely scrapped and replaced

with one running at around 1 MHz I expected to be buried in email, but I've

only received one message referring to this so far, and it wasn't from

anyone working on CAN-Do.

The CAN-Do module is unique in that it is going to be an integral component

INSIDE of every payload, so any deficiencies it may have are going to have a

large impact.

Let me restate my finding so far:

There are 4 categories of EMI and the CAN-Do module / 70 cm Receiver

combination is experiencing all four. They break down into radiated and

conducted emissions and susceptibility, and they are generally referred to

by a two-letter designation - RE, RS, CE, and CS.

RE and RS go together - the CAN-Do module's switching power supply inductor

radiates the 5 kHz switching noise out the back directly towards the

receiver (RE.) The Receiver's VCO's are both very sensitive to EMI and are

impacted by the CAN-Do module if they are within 4-1/2 inches of the

inductor (RS.) I've had to move the CAN-Do module off of the receiver PCB

and interconnect it with a ribbon cable to deal with this problem. The good

news is that I am fairly confident that It can be fixed by going to a

two-compartment chassis, with a bulkhead separating the CAN-Do module from

the analog Receiver to provide shielding. The bad news is that I think this

means we need a milled module chassis.

CE and CS also go together, and this is the real problem I see since you

can't fix conducted EMI with a shielded enclosure. It requires filtering.

The CAN-Do module is trashing the DC input from the power source and also

feeding noisy power to the Receiver. The outgoing noise is the bigger

concern because it will add to the CS problems for everything connected to

the power source. In the other direction, the switched power from the

CAN-Do module shows up in the IF output as 5 kHz spurs. Moving the CAN-Do

module physically away from the Receiver only dealt with the RE/RS issue. I

had to bypass the CAN-Do module and run clean power directly from the lab

bench supply to deal with the CE/CS problem. This means that there is no

current monitoring and no power control.

The 5 kHz switching frequency is bad for two reasons - it makes filtering

this noise a much larger problem than it needs to be, and the impact is more

severe since it is putting spurs all over the passband of the IF at 5 kHz

intervals.

If you sit back and think about the impact of a dozen noisy power supplies

all feeding EMI back to the common power source where they all mix together

and make their way back to each payload, it starts to look nasty. All these

supplies will be drifting around and beating with each other to produce sum

and difference noise on the power bus. 5 kHz noise is hard enough to get

rid of but what if there are difference components at a few hundred Hz? How

can you design a filter when you won't know what to expect until you hook

everything up and turn it on? And by then you're out of time.

I'm not making this stuff up. People I know have run into this exact

problem before and the result was very bad.

A while ago Howard Long made a suggestion that I think has great potential.

Here's what he had to say:

...in the original SDX PSU design I had in San Francisco last year is an

SMPS using the LM2672 device. These can be fitted with an AC coupled SYNC

signal to override the internal default SMPS frequency. I selected 375kHz

for my unit (6MHz divided by 16) to ensure its harmonics were outside the

10.7MHz IF passband. If the external SYNC fails the internal SMPS oscillator

takes over.

My Suggestions:

1) Revise the CAN-Do module to move the switching frequency up as far as

possible to move spurs out of the passband of sensitive analog circuitry,

and to ease the burden on EMI filtering.

2) The power and grounds must be filtered in both directions to minimize CE

back to the power source and to the payload.

3) The switching inductor should be a shielded to reduce RE inside the

module chassis.

4) A power supply capable of synching to a master oscillator should be

strongly considered.

This topic needs to be elevated to the top of the queue. The EMI

environment surrounding the CAN-Do module impacts the design of the next

revision of the 70 cm Receiver, and also directly impacts the chassis

design. What do we need to do to get going on this?

73,

Juan - WA6HTP

_______________________________________________

Via the Eagle mailing list courtesy of AMSAT-NA

[email protected]

http://amsat.org/mailman/listinfo/eagle

_____

_______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

Lou,

Thanks for sending that to me. Have any prototypes been tested of the converter? I didn't see anything about the switching frequencies. Where will they fall? Based on what you have, could an estimate be made as to the level and characteristics of conducted EMI on the bus? That could become the basis for an EMI requirement for the payloads. Then, if some reasonable method can be devised to create that conducted EMI, we'll all be in a position to test the payload designs against that EMI requirement, using the test fixture. That would be a great thing!

73,

Juan

_____

From: Louis McFadin [mailto:[email protected]] Sent: Saturday, June 30, 2007 2:18 PM To: [email protected] Cc: 'John B. Stephensen'; 'Dave Black (Work)'; 'Dave Black (Home)'; 'David Smith'; [email protected]; '[email protected] Com'; 'Juan.Rivera (Work)' Subject: Re: U-band receiver changes to reduce EMI and improve thermal regulation

For a look at what is planned for the Eagle Power system, look at the presentation I prepared for Dayton this year. You can see it at "http://homepage.mac.com/w5did" Look in the w5did Eagle folder. The Max Solar Converter is a preliminary document annd not to be published but it describes the converter very well.

Lou McFadin

W5DID

[email protected]

On Jun 30, 2007, at 7:05 AM, Juan Rivera wrote:

Juan

Hi John,

I had a chat or two with Bob Davis about chassis and I think you can bring SMA connectors out one side and the CAN-Do out the end, but nothing out the back.

Given the amount of conducted 5 kHz noise I'm seeing, do you think you can filter it out? Don't forget, I'm feeding the receiver with bypassed clean DC from the bench supply at the moment. The CAN-Do switched DC is nasty.

In your next version why not raise the switching frequency up and get it out of the passband completely? What do you think of Howard's idea of synching the switching supplies?

Juan

Jim,

To help with your question, some good phase noise numbers to ponder are as follows:

Using direct synthesis techniques (multiply/mix/divide and not using any PLLs or VCO's - this would be like the Comstron and PTS series of synthesizers), which use a low noise 10 MHz reference and will provide phase noise performance as follows:

600 MHz - -138 dBc/Hz at 10 KHz offset 1.2 GHz - 132 dBc/Hz at 10 KHz and following the 6 dB increase each time the frequency is doubled, -114 dBc/Hz at 10 KHz at 9.2 GHz.

A very low noise VHF crystal controlled oscillator/multiplier designed for a X-Band military doppler RADAR will do -118 to -124 dBc/Hz at 10 KHz.

I have seen -132 dBC/Hz at 10 KHz at X-Band in the laboratory using a fundamental mode cylindrical cavity oscillator (about the size of a coffee can). That's close to the ultimate low noise X-Band source - the double cavity klystron. Some chaps in Australia are doing about that with something called a smaller whisper mode cavity - if memory serves me right.

A really good free running transistor, not FET, DRO at X-Band will do about 85 to 90 dBc/Hz at 10 KHz. The FET DRO will be 10 to 15 dB worse.

The point I'm trying to make is that these numbers are obtained with considerable pain (complexity and size). So to get -110 dBC/Hz at 10 KHz at C Band for instance, you'll have to resort to a sophisticated low noise VHF crystal oscillator locked to a low noise 10 MHz oscillator and then multiplied to C band using pretty tight filtering along the way to reduced the spurious to less than -60 dB. Not exactly your straight forward PLL IC chip and a good dielectric resonator VCO.

Now if you ask what phase noise and at what distance from the carrier is required for a particular modulation format, then those numbers can become a little more argumentative since many variable factors enter into the equation. It can also matter if the signal is used as a receiver LO or a transmitter LO.

Now finally - - - for my opinion about your question for what do you need for X band terrestrial narrow band work, it appears to me that most X-Band EME is done with LO's running around -90 to -110 dBc/Hz at 10 KHz which is what the typical CMI or Frequency West phase locked cavity oscillator/multipliers will do but I'm convinced even -70 dBc to -80 dBc/Hz at 10 KHz. (or even worse) will also work. More important, for CW and SSB, what is the phase noise at 0.5 to 4 KHz.

As a point of note, the phase noise functional specifications for the JPL Mars Reconnaissance Orbiter SDST (Small Deep Space Transponder) receiver and exciter are ≤ −20 dBc/Hz at 1 Hz, ≤ −60 dBc/Hz at 100–1000 Hz, and ≤ −70 dBc/Hz at 1–100 kHz.

Well hopefully - - - the satellite system guys have our required numbers figured out and will be sharing them with us sub-system component guys real soon!!!

Jim Sanford wrote:

Bill: -110 dbc/hz at 10khz offset is what I REMEMBER in my response to the question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?" My memory may very well be WRONG, which was the point of asking my question.

I agree with you, looks like -90dbC at 10 KHz, but given what Juan's already said about his sources, and the uncertainties of the SDR-IQ, that may be VERY good!

Do you have a different answer to my original question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?" If so, I'd like to engage in that conversation, both for Eagle and for my terrestrial microwave station.

I agree with your assessment of Juan's work -- ABSOLUTELY EXEMPLARY!!!!!!

Thanks & 73, Jim [email protected]

Bill Ress wrote:

...

Jim,

Where did the -110 dBC/Hz at 10 KHz come from? And at what frequency? That's pretty aggressive using straight forward VCO's and PLLs. From Juan's data with his snappy SDRIQ, it looks like that the 70 cm is doing about -90 dBc/Hz at10 KHz. I thought that's what I read from his early HP8566B dtata.

My straight forward S2 RX downconverter PLL LO does -100 dBc/Hz at 10 KHz.

We'd better get a good read on this since, if real, it forces more sophisticated LO designs.

Great work Juan!!

Regards...Bill - N6GHz

Jim Sanford wrote:

...

I think I remember Bob's answer as -110dbc at 10 KHz offset.

Thanks & 73, Jim [email protected]

Thank you all for this great conversation. I want to remind everyone that we need long term coherence in the receiver, which can only be accomplished by low phase noise, not for the purity of the tone needed to be heard by a human ear, but by the need to correlate for long known symbols to recover frequency and timing in the SMS text messaging system and to do COHERENT detection of the symbols transmitted. We want the analog transponder to support this. This is also contributes to the requirement for a much larger dynamic range in the system than we have had heretofore on our spacecraft.

This system needed was added to the requirements for the receiver after the initial requirements but not before the design was done and certainly not before it was built.

73's Bob N4HY

Grant Hodgson wrote:

Jim Sanford wrote:

...

Do you have a different answer to my original question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?" If so, I'd like to engage in that conversation, both for Eagle and for my terrestrial microwave station.

Needless to say, there is no right and wrong, it's very subjective. However, there are a couple of factors that can be used in calculating phase noise requirements.

For analogue systems, i.e. good old CW and SSB, there are two basic phase noise requirements for the local oscillators.

The first is that the LO needs to produce a tone that sounds 'clean' when listened to on an analogue receiver. The term T9 is often used, but is rarely defined. The parameter of Residual FM can be used to specify the cleanliness of a tone - it's subjective, but a value of 20Hz RMS IMHO is more than good enough for narrow-band voice communications. As it's a noise measurement, the bandwidth needs to be specified, and so 300Hz - 3kHz is appropriate for SSB. (Note that FM is different as much wider bandwidths are needed).

So, I would suggest a residual FM spec. of 20Hz from 300Hz-3kHz. This can be achieved with a 'flat' phase noise performance of -76dBc/Hz at offsets from 300Hz to 3000Hz, for example.

Residual FM can be measured directly, but not easily, or it can be determined from an integration of phase noise over the required bandwidth. A Google search will give the maths of converting phase noise to FM, I'll post a link if required. Software such as that from KE5FX has the conversion built in. Note that for the purposes of determining residual FM, it doesn't really matter what the shape of the phase noise plot looks like, it is the total area under the curve that counts. Also, the value of 40Hz is based on the response of the human ear - it is not based on the actual frequency of the local oscillator. Therefore, a 10GHz local oscillator with 40Hz FM will sound just as good as one with 40Hz FM at 1.8MHz. However, maintaining the required level of FM at microwave frequencies obviously becomes more challenging as the frequency is increased.

The second requirement is that for the phase noise spec. applies to reciprocal mixing; the process whereby a weak signal is buried in the noise of a strong signal which is at a different frequency. For the purposes of the Eagle analogue receivers, consideration needs to be given to the relative strength of any strong signals. The SAW filters will do a good job of reducing out of band signals, so we are left with the consideration of strong in-band signals.

For satellite work, the levels of the incoming signals should be much closer together than would be the case for terrestrial work. If we assume that the strongest signals will be 30dB higher than the weakest, at an offset of 5kHz, and that we don't want the strong signal to degrade the S/N of the weak signal by more than 3dB, then we can say that the phase noise requirement at 5kHz offset would be approximately -30 - log(2.7kHz) = -64dBc/Hz.

That might seem surprising, but it is based on a relatively narrow range of input signals. It assumes that the uplink signals are perfectly clean with no phase noise, and ignores radar interference. If the range of input signals is larger, then the phase noise needs to be better than this. Terrestrial stations need much better performance than this as they have to deal with a much larger range of input levels. I've got a feeling that this subject has been anlaysed before, but I couldn't find a mention of it whilst browsing EagleP.

regards

Grant _______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

Bob,

Could you help me relate "long term coherence" with "low phase noise" as it applies to the LO requirements, say for the U, S2 and C LO's and the modulation formats we are using?

Also we talk around it, but have we fixed the LO phase noise requirements starting with the 10 MHz satellite reference?

Regards...Bill - N6GHz

Robert McGwier wrote:

Thank you all for this great conversation. I want to remind everyone that we need long term coherence in the receiver, which can only be accomplished by low phase noise, not for the purity of the tone needed to be heard by a human ear, but by the need to correlate for long known symbols to recover frequency and timing in the SMS text messaging system and to do COHERENT detection of the symbols transmitted. We want the analog transponder to support this. This is also contributes to the requirement for a much larger dynamic range in the system than we have had heretofore on our spacecraft.

This system needed was added to the requirements for the receiver after the initial requirements but not before the design was done and certainly not before it was built.

73's Bob N4HY

Grant Hodgson wrote:

...

Jim Sanford wrote:

...

Do you have a different answer to my original question, "What is a good phase noise number to shoot for in a microwave narrowband (ssb/cw) system?" If so, I'd like to engage in that conversation, both for Eagle and for my terrestrial microwave station.

Needless to say, there is no right and wrong, it's very subjective. However, there are a couple of factors that can be used in calculating phase noise requirements.

For analogue systems, i.e. good old CW and SSB, there are two basic phase noise requirements for the local oscillators.

The first is that the LO needs to produce a tone that sounds 'clean' when listened to on an analogue receiver. The term T9 is often used, but is rarely defined. The parameter of Residual FM can be used to specify the cleanliness of a tone - it's subjective, but a value of 20Hz RMS IMHO is more than good enough for narrow-band voice communications. As it's a noise measurement, the bandwidth needs to be specified, and so 300Hz - 3kHz is appropriate for SSB. (Note that FM is different as much wider bandwidths are needed).

So, I would suggest a residual FM spec. of 20Hz from 300Hz-3kHz. This can be achieved with a 'flat' phase noise performance of -76dBc/Hz at offsets from 300Hz to 3000Hz, for example.

Residual FM can be measured directly, but not easily, or it can be determined from an integration of phase noise over the required bandwidth. A Google search will give the maths of converting phase noise to FM, I'll post a link if required. Software such as that from KE5FX has the conversion built in. Note that for the purposes of determining residual FM, it doesn't really matter what the shape of the phase noise plot looks like, it is the total area under the curve that counts. Also, the value of 40Hz is based on the response of the human ear - it is not based on the actual frequency of the local oscillator. Therefore, a 10GHz local oscillator with 40Hz FM will sound just as good as one with 40Hz FM at 1.8MHz. However, maintaining the required level of FM at microwave frequencies obviously becomes more challenging as the frequency is increased.

The second requirement is that for the phase noise spec. applies to reciprocal mixing; the process whereby a weak signal is buried in the noise of a strong signal which is at a different frequency. For the purposes of the Eagle analogue receivers, consideration needs to be given to the relative strength of any strong signals. The SAW filters will do a good job of reducing out of band signals, so we are left with the consideration of strong in-band signals.

For satellite work, the levels of the incoming signals should be much closer together than would be the case for terrestrial work. If we assume that the strongest signals will be 30dB higher than the weakest, at an offset of 5kHz, and that we don't want the strong signal to degrade the S/N of the weak signal by more than 3dB, then we can say that the phase noise requirement at 5kHz offset would be approximately -30 - log(2.7kHz) = -64dBc/Hz.

That might seem surprising, but it is based on a relatively narrow range of input signals. It assumes that the uplink signals are perfectly clean with no phase noise, and ignores radar interference. If the range of input signals is larger, then the phase noise needs to be better than this. Terrestrial stations need much better performance than this as they have to deal with a much larger range of input levels. I've got a feeling that this subject has been anlaysed before, but I couldn't find a mention of it whilst browsing EagleP.

regards

Grant _______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle

Folks,

Don't forget that the 70 cm receiver also has an internal reference oscillator than can be selected, at least in this current version. The phase noise of both oscillators need to be considered.

Juan WA6HTP

On 6/11/07, Matt Ettus [email protected] wrote:

Robert McGwier wrote:

...

Thank you all for this great conversation. I want to remind everyone that we need long term coherence in the receiver, which can only be accomplished by low phase noise, not for the purity of the tone needed to be heard by a human ear, but by the need to correlate for long known symbols to recover frequency and timing in the SMS text messaging system and to do COHERENT detection of the symbols transmitted. We want the analog transponder to support this. This is also contributes to the requirement for a much larger dynamic range in the system than we have had heretofore on our spacecraft.

This will be completely controlled by the master reference oscillator. Are there specs for this? Is someone designing one?

Matt

---

Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle