The Boonton is crystal stabilized. it works quite well. There is software in it + a counter to make that all work. Some type of software PLL working with the counter which I think is reciprocal. All referenced to a 10 MHz standard (something I would push for). One analog multiplier is devoted to frequency trim between steps. The algorithm probably lifted from RF synths.
As for cancelling techniques RCA beat us to it in 1939. More stuff here: BroadcastHistory - RCA
Attachments
To measure an ultra pure DUT you do not need an ultra pure generator and an ultra pure digitizer. It is enough that generator and digitizer have zero error together and that can be achieved by trimming. There is a lot that can be done in the lookup tables for a DDS with a 500 MHz 16 bit dac or in the post processing of a 200 MSPS 16 bit ADC. It might be possible in the analogue domain also, but more clumsy.
There are limits, of course. When the generator has 10% harmonics, that may change the behaviour of the DUT. But when we are already at -130 dBc without matching, second order errors will be down at -260 or play no role at least.
Also, the gain of the DUT would have to be 1, or we would have to accept at least the distortion of a precision attenuator.
regards, Gerhard
There are limits, of course. When the generator has 10% harmonics, that may change the behaviour of the DUT. But when we are already at -130 dBc without matching, second order errors will be down at -260 or play no role at least.
Also, the gain of the DUT would have to be 1, or we would have to accept at least the distortion of a precision attenuator.
regards, Gerhard
Hi Demian,
-Chris
Assuming you have the equipment that can be referenced to a standard. In this case I do and would also recommend others to do the same. You can then pickup equipment without the high stability oscillator because it is replaced by a good standard. Many GPS timing systems do have a 10 MHz output, some also have 5 MHz and 15 MHz outputs.
HP bought Boonton and continued production of that line. Not surprising when technology is shared between companies. That's the HP way!
-Chris
Haven't made any claims with the idea.
Of course the print is too small for me to read.
Last edited:
Well I did just that but haven't used it much yet. I talked the guy down in price. I waited a long time before the item was shipped and the guy made sure he sent me one with a really noisy fan. Still have to fix that.
Just need the GPS unit now.
Yes I have looked at the Boonton. The schematic is spread over many pages and is hard to follow. I think much is done in the micro hidden in the code.
That depends on what the equipment does. Long time stability is easy, GPS may bring you to Cs levels if you have good sat visibility.
Short term stability / phase noise is a different story. That is determined by the oscillator only. I decided to pay a few hundred premium for the Wenzel oven in my SR620 counter, in spite of the GPS receivers that share the same rack.
My GPSDO came with a Wenzel. But that's not common. And for audio circuits probably not relevant.
The Boonton circuit is spread across three boards and the disciplining is done in the micro. You may be able to look at the code used for GPSDO's. I believe there are some DIY and open source projects for making GPSDO's.
The Boonton circuit is spread across three boards and the disciplining is done in the micro. You may be able to look at the code used for GPSDO's. I believe there are some DIY and open source projects for making GPSDO's.
The catalog is downloadable from the site. For some reason its protected so not printable. All I could do is a screen grab. The device has several controls to adjust gain and phase to cancel out the signal from the DUT with an inverted copy.
back tracking a bit... concept.....
a F to V converter (DC) driving a pot(s) which forms the R part of a varying freq filter. Doesnt have to be precision freq - not digital... just an octave above fundamental freq ( 2H and >) Can be switched in range with same circuit which switches the freq gen.
Does that spur any ideas?
THx-RNMarsh
a F to V converter (DC) driving a pot(s) which forms the R part of a varying freq filter. Doesnt have to be precision freq - not digital... just an octave above fundamental freq ( 2H and >) Can be switched in range with same circuit which switches the freq gen.
Does that spur any ideas?
THx-RNMarsh
I agree that analog processing is likely the best for ultra-pure. As far as frequency accuracy and stability, gentle, non-intrusive phase-locked loop techniques can solve that with little or no degradation to signal purity of the oscillator. The loop can lock the oscillator to a highly stable reference, whether it be a more stable analog reference or a frequency-synthesized reference. This applies in the most practical sense to an SVO.
Bear in mind that the THD analyzer that uses SVF for its notch must "lock" to the incoming fundamental, and this is really little more than a phase locked loop with a fully synchronous phase detector. This is how mine and many other THD analyzers work. The fact that this "phase-locked loop" SVF can implement an analyzer with very, very low residual is testament that, if well designed, it can be used with equally low degradation to purity to lock the oscillator to a precision reference.
Things get a little more complicated if additional non-SVF-based filter elements are also used, such as twin-T notch filters.
Cheers,
Bob
Good point. In fact, although it is a different approach, my distortion magnifier (DM) also has this property - e.g. that it greatly reduces the degradation of the measurement caused by oscillator harmonics - approximately by the amount of the distortion magnification.
Cheers,
Bob
The "diamond bridge" is the basic principle, Bob have you ever thought of an auto-tuned, maybe an R/C ladder and relays version of DM?
https://archive.org/details/philtrans06307436
Hi Gerhard,
Hi David,
The first system I bought was called a Trimble T-Bolt, or Thunderbolt. It does have an internal oven oscillator and 10 MHz output, as well as a pulse per second output. I am building a receiver / oscillator from a PCB I bought through "Time Nuts", a yahoo group. I'm adding a Symetricom (bought by Microsemi) receiver / oscillator to the system. This is all being shared through an HP 5087 distribution amplifier. Still need some more cards for that one. This amplifier uses tuned channels to filter out harmonics or other frequencies you haven't tuned it for. Some day I plan on building a distribution amplifier with more channels and higher performance.
Long story short, complete GPS disciplined oscillators can be found on Ebay. $200 seems to be a starting price for a system that includes the power supply and timing antenna (different from a standard GPS location type antenna). You just need some cable (RG-58 or better depending on length and signal strength). So while these systems aren't cheap, they aren't extremely expensive if you shop carefully.
You can even use the high stability oscillator option in one of your pieces of equipment and distribute that to everything else. Later you can add a GPS receiver if you want.
-Chris
The lowest phase noise will be with a quartz crystal. So if you need extreme low jitter performance, put the GPS receiver in hold-over mode to stop it from disciplining the oscillator for the duration of your test. My GPS receivers came with double oven x-tal oscillators designed for very low phase noise. I forget what they call how these crystals are cut, SC I think. Performance exceeds the normal crystal oscillators.
Hi David,
The first system I bought was called a Trimble T-Bolt, or Thunderbolt. It does have an internal oven oscillator and 10 MHz output, as well as a pulse per second output. I am building a receiver / oscillator from a PCB I bought through "Time Nuts", a yahoo group. I'm adding a Symetricom (bought by Microsemi) receiver / oscillator to the system. This is all being shared through an HP 5087 distribution amplifier. Still need some more cards for that one. This amplifier uses tuned channels to filter out harmonics or other frequencies you haven't tuned it for. Some day I plan on building a distribution amplifier with more channels and higher performance.
Long story short, complete GPS disciplined oscillators can be found on Ebay. $200 seems to be a starting price for a system that includes the power supply and timing antenna (different from a standard GPS location type antenna). You just need some cable (RG-58 or better depending on length and signal strength). So while these systems aren't cheap, they aren't extremely expensive if you shop carefully.
You can even use the high stability oscillator option in one of your pieces of equipment and distribute that to everything else. Later you can add a GPS receiver if you want.
-Chris
Last edited:
Any distortion compensation technique works well, when the nonlinearity, or distortions phases and levels are stable, but nonlinearity in real devices may drift. This may give a real pain, when the compensation must be done at very low distortion levels.
Victor.
My HP universal counter didn't come with the C option so I would have to have an external reference or live with what I have.
The model number eludes me right now.
No, I haven't really thought of doing that, but it sure would be convenient if the auto-tune circuitry could be made in such a way that it would not compromise distortion.
Just as in a THD analyzer, two variable must be auto-tuned using quadrature versions of the error. First, and most obviously, the amplitude for nulling. Secondly, however, there needs to be some adjustment of phase for higher-frequency measurements where the HF rolloff of the amplifier under test causes some phase lag.
Cheers,
Bob