Cheap OCXO as external clock for HackRF, Airspy R2 and similar SDRs

[Ayushman Tripathi]

Hi,

I managed to get the external clock working on my HackRF One using a 12 USD OCXO from Banggood. The issue was these cheap OCXOs output 5V TTL, while the HackRF and other SDRs like the Airspy R2 expect a 3.3V clock input, so connecting it directly will not work and can damage the SDR/CLK chip. The absolute max on the HackRF CLKIN is 3.8V for the Si5351 clock chip inside.

My solution was a 10 dB SMA attenuator in between. Since CLKIN is high impedance, the pad only attenuates around 5 dB instead of 10, which lands the 5V TTL right in the valid window.

HackRF successfully detects it: 
ayushman@ayushman-i914900KS:/mnt/extra$ hackrf_clock -i
CLKIN status: clock signal detected
ayushman@ayushman-i914900KS:/mnt/extra$ hackrf_clock -i
CLKIN status: clock signal detected

I jerry-rigged the scope connection with 2 bare wires poked into the BNC, as my SMA to BNC adapter arrives Monday, so the attached waveform shots are just a quick look showing a 10 Mhz square wave. I'll post proper measurements next week.

If you try this: do not use the sine output of the OCXO, it swings negative, and the Si5351 does not tolerate below -0.5V.

Ayushman

[Ayushman Tripathi]

These 16 USD OCXO boards on Amazon look identical to the one I got from Banggood, same board layout with TTL and sine outputs, so I think they are the same unit: 

https://www.amazon.com/Frequency-Oscillator-Temperature-Instrument-benchmarks/dp/B0BBVTPPHZ

[Stephen Arbogast]

Very  Nice!

I  still have my  first  dish, WiFi from Amazon, and  the  GPIO  Labs lna  I used   for  my first HI detection and  a  GPSDO.   I  think  I 'm going to  need to buy  another AirSpy R2  and get set up for  some  correlation.   Marcus  Leech of  BAA  wrote  a GNU Radio flow graph  to do this..

See...

https://gpio.com/products/hydrogen-line-lna

https://www.leobodnar.com/shop/index.php?main_page=index&cPath=107

https://github.com/ccera-astro/baa_seminar

I  remember  plugging my GPSDO  into  my  HackRF-One   for  a GNSS application...  really  helped.

Stephen

[Ayushman Tripathi]

Thanks, Stephen!

Yes, I will get the Leo Bodnar LBE1420 mini GPSDO you recommended for my narrowband experiments. Right now the Airspy Mini's internal clock drifts around 0.05 to 0.1 Hz per second at 1420 MHz, which is about the same as the 0.15 Hz/s sidereal drift I am trying to measure, you can see the details in my narrowband pipeline post. So a proper GPSDO is the real fix when I get an R2.

I know I will probably never find any signal, the odds of detecting a signal like Wow! (whether natural or artificial) are extraordinarily low. But building a pipeline that can actually do the search properly is worth it on its own, and the search is interesting :)

It's also good to hear your GPSDO into the HackRF worked well for GNSS. And thanks for the Marcus Leech flow graph link, will study it for the interferometry idea which I was also thinking about, but postponing it for a while.

Ayushman

[Don Latham]

They are indeed the same. Also, the one I bought could not be brought to 10 MHz with the adjusting pot, comparing to a GPS governed source.  If accuracy is necessary, a GPS source is much better.  Outside of that or even a Rb source. set the adjustable source to WWV by careful zero beat.

There are some GPS sources for around $70, although adjustment is done and then set rather than continuous guidance by satellite.

------------
Don Latham
PO Box 404, guidance 
Frenchtown, MT, 59846
406-626-4304

 

---

From: Ayushman Tripathi <ayushmantr...@gmail.com>
To: Society of Amateur Radio Astronomers <sara...@googlegroups.com>
Date: Saturday, 18 July 2026 3:53 PM MDT
Subject: [SARA] Re: Cheap OCXO as external clock for HackRF, Airspy R2 and similar SDRs

--
--
You received this message because you are subscribed to the Google
Groups "Society of Amateur Radio Astronomers" group.
To post to this group, send email to sara...@googlegroups.com
To unsubscribe from this group, send email to
sara-list-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sara-list?hl=en
---
You received this message because you are subscribed to the Google Groups "Society of Amateur Radio Astronomers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sara-list+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sara-list/b070ddca-4d05-4312-8f58-b7a7c490850cn%40googlegroups.com.

[Ayushman Tripathi]

Hi,

The SMA to BNC adapter arrived early, here are the measurements I did now using an oscilloscope. Numbers differ a bit from the attached plots because the scope needed a calibration correction, which the final figures account for.

The TTL output is about 4.3V peak, so it is a real 5V-family TTL driver, the spec sheet says nothing about the TTL output at all. The duty cycle is around 60%, which explains the multimeter reading, 4.3 x 0.6 is about 2.6V average, exactly what the meter shows.

Through the 10 dB attenuator about 1.6V reaches CLKIN. Below the Si5351's 2.31V datasheet minimum, so on paper it should not detect, but it detects and locks reliably anyway. Given the raw output is 4.3V, do not connect directly, use at least a 5 dB attenuator. 10 dB works fine, as mine shows.

The sine output is close to spec but swings down to about -1.6V, far past the -0.5V absolute minimum, see the attached image. Never use the sine output for CLKIN for HackRF.

Plots attached.

On Saturday, July 18, 2026 at 5:40:06 PM UTC-4 Ayushman Tripathi wrote:

[Ayushman Tripathi]

Hi Don,

Thanks for the info. By any chance do you mean this type of GPSDO?

https://www.banggood.com/GPSDO-GPS-Tamed-Thermostatic-Crystal-Oscillator-GPS-Tamed-Clock-10MHz-Signal-Source-Global-Positioning-System-Disciplined-Oscillator-p-2009155.html

That one is around 70 USD, so I am wondering if it is the same class of unit you had in mind.

[Don Latham]

yes. These generate 1 second ticks from the gps and from the local oscillator. The local oscillator is then adjusted to the gps by setting its voltage (OXCO) using the time difference. User is supposed to "freeze" the set voltage after some settling time. I have observed the 10 MHZ and 1 pps signals on a scope and the adjustment in operation can be seen as jitter. I have no idea what the statistics look like. But the price is right. To avoid problems using this beast, leave it adjusting until beginning datataking and then freeze the crystal. I'm open to argument on this point.

 

------------
Don Latham
PO Box 404,

Frenchtown, MT, 59846
406-626-4304

 

---

From: Ayushman Tripathi <ayushmantr...@gmail.com>
To: Society of Amateur Radio Astronomers <sara...@googlegroups.com>

Date: Sunday, 19 July 2026 7:30 PM MDT
Subject: Re: [SARA] Re: Cheap OCXO as external clock for HackRF, Airspy R2 and similar SDRs

To view this discussion visit https://groups.google.com/d/msgid/sara-list/4a61b33c-d580-4b46-bc1a-a38e22f4e7b2n%40googlegroups.com.

[Ayushman Tripathi]

Hi Don,

Thanks, that makes sense. I will get one of these GPSDOs soon and put it on the scope once it arrives, will share the results here.

[Don Latham]

I would also have a good look at Whit Reeve's study. Lots of cogent information.

 

------------
Don Latham
PO Box 404,
Frenchtown, MT, 59846
406-626-4304

 

---

From: Ayushman Tripathi <ayushmantr...@gmail.com>
To: Society of Amateur Radio Astronomers <sara...@googlegroups.com>

Date: Monday, 20 July 2026 4:56 PM MDT

To view this discussion visit https://groups.google.com/d/msgid/sara-list/b0428822-2c26-48c8-a267-1bcca4631000n%40googlegroups.com.