|
I stumbled across a bunch of precision OXCO references on ebay for cheap. These are new boards made up using used CTI OXCOs from obsolete cell tower kit. The following is one of many sellers. The title is the search key to use.
The OXCOs are marked
CTI 140503
10.000MHz
OSC5AB02
I've been comparing a pair to a GPSDO and they are easily on a par with the high accuracy HP time bases and easier to adjust. If made up in a box with a BNC connector they are a great way to upgrade the accuracy of lots of HPAK kit without needing an internal modification. Just plug and select external reference. Naturally the ideal fix for any dead 10 MHz timebase oscillator. They run on 7-13 V and at 12 V consume 250 mA. More if the oven isn't insulated.
Yellow trace is GPSDO. Blue trace is hand trimmed module. Purple trace is the OXCO in my 5386A trimmed as best I could. Thermal effects are probably the biggest issue in the frequency variation.
One of the two ebay units I adjusted was reading 0.050 Hz high with my 5386A locked to the GPSDO when I first connected it. It's been running for 2 days or more. After I put it back between some double scrap bath towel the error has dropped to <0.01 Hz.
Have Fun!
Reg
|
"Reginald Beardsley via groups.io" <pulaskite@...> writes: I stumbled across a bunch of precision OXCO references on ebay for cheap. These are new boards made up using used CTI OXCOs from obsolete cell tower kit. The following is one of many sellers. The title is the search key to use.
The OXCOs are marked
CTI 140503
10.000MHz
OSC5AB02
I've been comparing a pair to a GPSDO and they are easily on a par with the high accuracy HP time bases and easier to adjust. If made up in a box with a BNC connector they are a great way to upgrade the accuracy of lots of HPAK kit without needing an internal modification. Just plug and select external reference. Naturally the ideal fix for any dead 10 MHz timebase oscillator. They run on 7-13 V and at 12 V consume 250 mA. More if the oven isn't insulated.
Yellow trace is GPSDO. Blue trace is hand trimmed module. Purple trace is the OXCO in my 5386A trimmed as best I could. Thermal effects are probably the biggest issue in the frequency variation.
One of the two ebay units I adjusted was reading 0.050 Hz high with my 5386A locked to the GPSDO when I first connected it. It's been running for 2 days or more. After I put it back between some double scrap bath towel the error has dropped to <0.01 Hz. Any idea how it performs in regard to phase noise? Matt
|
What is the physical size?
"Reginald Beardsley via " <pulaskite=[email protected]>
writes:
> I stumbled across a bunch of precision OXCO references on ebay for cheap.? These are new boards made up using used CTI OXCOs from obsolete cell tower kit.? The following is one of many sellers. The title is the search key to use.
>
>
>
> The OXCOs are marked?
>
> CTI 140503
> 10.000MHz
> OSC5AB02
>
> I've been comparing a pair to a GPSDO and they are easily on a par with the high accuracy HP time bases and easier to adjust.? ? If made up in a box with a BNC connector they are a great way to upgrade the accuracy of lots of HPAK kit without needing an internal modification. Just plug and select external reference.? Naturally the ideal fix for any dead 10 MHz? timebase oscillator.? They run on 7-13 V and at 12 V consume 250 mA. More if the oven isn't insulated.
>
> Yellow trace is GPSDO.? Blue trace is hand trimmed module.? Purple trace is the OXCO in my 5386A trimmed as best I could.? Thermal effects are probably the biggest issue in the frequency variation.
>
> One of the two ebay units I adjusted was reading 0.050 Hz high with my 5386A locked to the GPSDO when I first connected it.? It's been running for 2 days or more.? After I put it back between some double scrap bath? towel the error has dropped to <0.01 Hz.
Any idea how it performs in regard to phase noise?
Matt
|
They appear to be Chinese made. The lowest price I saw for an
American seller was $20 with free shipping. If you check out
the url on the board, it takes you to a Chinese site with lots
of RF gadgets.
Looks like a nice deal.
Paul
toggle quoted message
Show quoted text
On Mon, Apr 10, 2023 at 05:16:38PM -0700, Reginald Beardsley via groups.io wrote: I stumbled across a bunch of precision OXCO references on ebay for cheap. These are new boards made up using used CTI OXCOs from obsolete cell tower kit. The following is one of many sellers. The title is the search key to use.
The OXCOs are marked
CTI 140503
10.000MHz
OSC5AB02
I've been comparing a pair to a GPSDO and they are easily on a par with the high accuracy HP time bases and easier to adjust. If made up in a box with a BNC connector they are a great way to upgrade the accuracy of lots of HPAK kit without needing an internal modification. Just plug and select external reference. Naturally the ideal fix for any dead 10 MHz timebase oscillator. They run on 7-13 V and at 12 V consume 250 mA. More if the oven isn't insulated.
Yellow trace is GPSDO. Blue trace is hand trimmed module. Purple trace is the OXCO in my 5386A trimmed as best I could. Thermal effects are probably the biggest issue in the frequency variation.
One of the two ebay units I adjusted was reading 0.050 Hz high with my 5386A locked to the GPSDO when I first connected it. It's been running for 2 days or more. After I put it back between some double scrap bath towel the error has dropped to <0.01 Hz.
Have Fun!
Reg
!DSPAM:6434a6f329911317811413!
--
Paul Amaranth, GCIH | Manchester MI, USA
Aurora Group of Michigan, LLC | Security, Systems & Software
paul@... | Unix/Linux - We don't do windows
|
The boards are 45 x 65 mm. I just ordered 3 more and probably cleaned out the USA shipper :-( Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO. That is 0.5 ppb! For $20? Wow! I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age, BTW they stabilize amazingly quickly. Much faster than the 5386A..
Reg
|
Thanks. I am thinking of an adapter board and cable to add option 001 to the HP3325A/B function generators. The adaptor would fit the mounting studs in the unit and supply power from the existing wiring. I have at least a half doven that may need his.
On Tue, Apr 11, 2023 at 9:25?AM Reginald Beardsley via <pulaskite= [email protected]> wrote: The boards are 45 x 65 mm.? I just ordered 3 more and probably cleaned out the USA shipper :-(? Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO.? That is 0.5 ppb!? For $20?? Wow!? I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age,? BTW they stabilize amazingly quickly.? Much faster than the 5386A..
Reg
|
An extremely similar, possibly the sane but the crystal marking is different, board is available from Aliexpress.
Free shipping to UK @ ?11.47 (approx $14) each.
I've bought 4, although if as good as Reg's I'll struggle to check them.
Michael Kellett,
MK Electronics Ltd.,
toggle quoted message
Show quoted text
On 11/04/2023 14:25, Reginald Beardsley via groups.io wrote: The boards are 45 x 65 mm. I just ordered 3 more and probably cleaned out the USA shipper :-( Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO. That is 0.5 ppb! For $20? Wow! I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age, BTW they stabilize amazingly quickly. Much faster than the 5386A..
Reg
|
?An extremely similar, possibly the same but the crystal marking is different, board is available from Aliexpress.
Free shipping to UK @ ?11.47 (approx $14) each.
I've bought 4, although if as good as Reg's I'll struggle to check them.
Michael Kellett,
MK Electronics Ltd.,
|
On 4/11/23 08:23, Michael A. Terrell wrote: I am thinking of an adapter board and cable to add option 001 to the HP3325A/B function generators. The adaptor would fit the mounting studs in the unit and supply power from the existing wiring. I have at least a half doven that may need his. If you are making your own board, there are bare ocxo's for sale on epay in batches of ten. ($30) Not same model, square wave only is the most common. Does the HP3325A/B want sine or square?
|
Am 11.04.2023 um 17:26 schrieb John Griessen: On 4/11/23 08:23, Michael A. Terrell wrote:
I am thinking of an adapter board and cable to add option 001 to the HP3325A/B function generators. The adaptor would fit the mounting studs in the unit and supply power from the existing wiring. I have at least a half doven that may need his. If you are making your own board, there are bare ocxo's for sale on epay in batches of ten. ($30)
Not same model, square wave only is the most common. The linked board from CN uses a square wave OCXO and filters it with a simple CLC lowpass? (C12,L2,C13) for "sine wave" output. And it seems it is directly coupled to the TTL out, so the lowpass directly distorts the rise/fall time of it. See and for some details of the OCXO. Bernd
|
So why are you bothering with OCXOs when you have a GPSDO, Reg?? Lower phase noise?? ? ? ? ?Jim Ford?
Sent from my T-Mobile 4G LTE Device
toggle quoted message
Show quoted text
-------- Original message -------- From: "Reginald Beardsley via groups.io" <pulaskite@...> Date: 4/11/23 6:25 AM (GMT-08:00) Subject: Re: [HP-Agilent-Keysight-equipment] Cheap and simple OXCOs upgrades on ebay
The boards are 45 x 65 mm.? I just ordered 3 more and probably cleaned out the USA shipper :-(? Sorry. FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO.? That is 0.5 ppb!? For $20?? Wow!? I'll try to set up to log via GPIB shortly. I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz. This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards. The thing about OXCOs is they just get better with age,? BTW they stabilize amazingly quickly.? Much faster than the 5386A.. Reg
|
Similar to the one I this evening transplanted into my newly acquired HP 5386A. The PCB design I found on EEVblog. The reason for the run of Teflon coax is because the PCB is originally intended for the 5384A and 5385A counter models.
Wilko
|
For best stability the board should use a precision reference or the internal reference in the OCXO if it has one.
The output from a 5V regulator is not stable enough over time and temperature.
|
Thank you! I'm using an Instek lab supply at 12 V drawing 250 mA. The Instek channel is rated to 32 V at 3 A. So I'm pretty sure power supply is not a major issue. But a very obvious and important consideration. An LM399 would be a huge improvement I'm sure. I have several I can press into service. So very helpful test suggestions.
However, I have enough bench experience to feel certain that thermal effects dominate what I'm seeing. So I'll pursue an LM399 later.
Precisely the reason I thought this worthy of this list. It's already 10x better than the high precision OXCO in my 5386A measured using a scope triggering on Leo's GPSDO.
BTW the *only viable way to adjust a precision reference is a scope!!!
The ebay gadgets are a turnkey item. One can always unsolder the 5 connections to the OXCO and install in a custom board. Ideal for a cheap GPSDO build using a U-Block. I strongly expect that the schematic is the datasheet eval board. That is the usual case I've observed when exotica shows up cheap. I just haven't gone datasheet hunting as my ISP is blocking dangerous sites such as elegoo.com :-( And I spent the day having 2 supervisors tell me that it won't change
Voltage and frequency references have exponential convergence to a terminal value with time under power. As a consequence long term (years) accuracy is excellent for devices which have been in continuous operation for tens of thousands of hours.
So far since I set the 5386A to use Leo's GPSDO, I've observed 0.1 ppb accuracy with the board in between two layers of Turkish towel top and bottom. It's hard to say which is more amazing, that I set it perfectly or that it's so stable.
If you are in the position that you can't justify the cost of a GPSDO, but know a local person who has one and is willing to help you trim it you'll have a superb reference. It puts out 2.4 Vpp into 50 ohms so a simple splitter would feed 2 and probably 3 instrument external reference inputs. A $30-40 upgrade of all the instruments on your bench with precision timing requirements to 0.1 ppb including the cabling seems like a hell of a deal to me.
Reg
|
These boards are pulls
from previously working systems so have a good bit of aging
already.
You can also buy just
the OXCO's, rather than packaged assemblies. Search eBay for
OSC5AB02. I bought 2 for 8.02.
John?? KK6IL
On 4/11/2023 6:25 AM, Reginald
Beardsley via groups.io wrote:
toggle quoted message
Show quoted text
The boards are 45 x 65 mm. I just ordered 3 more and probably cleaned out the USA shipper :-( Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO. That is 0.5 ppb! For $20? Wow! I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age, BTW they stabilize amazingly quickly. Much faster than the 5386A..
Reg
|
Dumb question but what is the Vcont pin for?
Thanks,
Barry - N4BUQ
From: "John Kolb" <jlkolb@...>
To: "HP-Agilent-Keysight-equipment" <[email protected]>
Sent: Tuesday, April 11, 2023 7:59:52 PM
Subject: Re: [HP-Agilent-Keysight-equipment] Cheap and simple OXCOs upgrades on ebay
These boards are pulls
from previously working systems so have a good bit of aging
already.
You can also buy just
the OXCO's, rather than packaged assemblies. Search eBay for
OSC5AB02. I bought 2 for 8.02.
John?? KK6IL
On 4/11/2023 6:25 AM, Reginald
Beardsley via groups.io wrote:
The boards are 45 x 65 mm. I just ordered 3 more and probably cleaned out the USA shipper :-( Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO. That is 0.5 ppb! For $20? Wow! I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age, BTW they stabilize amazingly quickly. Much faster than the 5386A..
Reg
|
Most likely it's the EFC (Electronic Frequency Control) pin.
On Tuesday, April 11, 2023 at 08:04:41 PM CDT, n4buq <n4buq@...> wrote:
Dumb question but what is the Vcont pin for?
Thanks,
Barry - N4BUQ
From: "John Kolb" <jlkolb@...>
To: "HP-Agilent-Keysight-equipment" <[email protected]>
Sent: Tuesday, April 11, 2023 7:59:52 PM
Subject: Re: [HP-Agilent-Keysight-equipment] Cheap and simple OXCOs upgrades on ebay
These boards are pulls
from previously working systems so have a good bit of aging
already.
You can also buy just
the OXCO's, rather than packaged assemblies. Search eBay for
OSC5AB02. I bought 2 for 8.02.
John?? KK6IL
On 4/11/2023 6:25 AM, Reginald
Beardsley via groups.io wrote:
The boards are 45 x 65 mm. I just ordered 3 more and probably cleaned out the USA shipper :-( Sorry.
FWIW the one I am testing is now varying +/- 0.01 Hz either side of 10 MHz against a Leo Bodnar GPSDO. That is 0.5 ppb! For $20? Wow! I'll try to set up to log via GPIB shortly.
I strongly believe that if placed in an outer oven with tight temperature control you could get to 0.001Hz.
This is a once in a lifetime deal like all the stuff that the RoHS idiocy dumped into the cheap ebay eval boards.
The thing about OXCOs is they just get better with age, BTW they stabilize amazingly quickly. Much faster than the 5386A..
Reg
|
Thank you John!
I just bought 10 bare modules for $30. Will 10 OXCOs synchronize by consensus at precisely 10 MHz? Inquiring minds need to know. I rather fear I'm in danger of becoming a time-nut.
Making replacement clocks for various HP kit looks as it it could be profitable for someone tooled up to make them with instrument specific board form factor PCBs. Not something I want to do at 70, but were I younger would likely do.
Have Fun!
Reg
|
On 4/11/23 21:49, Reginald Beardsley via groups.io wrote: I just bought 10 bare modules for $30. Will 10 OXCOs synchronize by consensus at precisely 10 MHz? Inquiring minds need to know. Injection locking? I rather fear I'm in danger of becoming a time-nut. It's a very slippery slope. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
I used an external sinewave reference with the 3325A, on my bench at Microdyne, but I think that it can use either.
A crystal filter will clean up the square wave better than the filter on these boards. I'm trying to schedule cataract surgery, and I simply can't wade through all of the related text in the manuals.
|
Michael A. Terrell
Dave McGuire