Rubbitin frequency standard arrived and running

Rubbitin frequency standard arrived and running:

formatting link
output 10MHz 1Vpp:
formatting link
As this is a 10MHz scope, it probably shows a sinewave even if it is not. Spectrum later.

I used a very cheap Toshiba 15V 3A laptop supply for the 15V: ebay 221153140002

The rubbitin guy, ebay 280655233263 was so nice as to include a 7805 in the package for the 5V, here it is still running on the lab supply, 4.9V

Reply to
Jan Panteltje
Loading thread data ...

formatting link

Two questions.

How did you verify the accuracy?

You have a device with partspergillion accuracy/stability. Why would you even consider powering it from a supply that adds parts per thousand of supply switching/ground noise? Switcher is better from a loss standpoint. But is that really your primary objective?

Reply to
mike

On a sunny day (Sat, 01 Dec 2012 11:14:43 -0800) it happened mike wrote in :

formatting link

It is my refernce, tehre are RS232 command to set a correction. Have not got around to testuing that soft, am building the little switch mode

5V regulator, now tested, needs soem hoel cuttin gin its box, has RS232 connector for PC..

Because it has nothing to do with that, did you read the specs?

fun

Reply to
Jan Panteltje

formatting link

formatting link

That's the point. It's a secondary standard of unknown calibration accuracy. What correction data are you gonna stuff in that RS-232 port?

Some people will look at part of the system and argue that the input is irrelevant. Others might consider the total system that they may implement now or later and decide that switching noise on the power supply or injected into ground might affect whatever is hooked to the reference output and choose to avoid complication by using the cleaner 7805. But that's just me.

You don't have to average out noise that you don't add.

Reply to
mike

On a sunny day (Sat, 01 Dec 2012 13:02:42 -0800) it happened mike wrote in :

No it is not, if you had followed my earlier posting on this subject, then you would have read I will take the 50.xxxxxxx MHz from the physics module directly and feed it in my FPGA board as replacement for the 50.xxx MHz crystal oscillator module. All other frequencies are derived from that on that board.

DID YOU READ THE SPECS????

These factors are given, and if you can do basic multiplication (not as in sex) and divivison, and measure ripple, you can CALCULATE the effect. Then if you can supply NUMBERS for the accuracy you want, and not senseless assumptions, you can see if it meets your specs.

Cleaner? Its a shit regulator with a large initial tolerance and bad supply rejection. It is used a lot to supply digital stuff that works from 4.75 to 5.25 V, for analog stuff I would avoid it, LM317 is way better. The 7805 has as plus low component count (add 2 caps), and foldback current limiting. It has as minus too that it gets freaking hot and likely needs a heatsink in many cases, and the case size here is 50x35x20mm plastic in this case (kees case etc), would melt or deform.

sigh.

Stay clear of them power drinks.

Reply to
Jan Panteltje

Rubbitin frequency standard arrived and running:

I am actually communicating with teh unit:

panteltje: ~/compile/pantel/fpga/jps # ./ft5680com -d /dev/ttyUSB0

Menu fENTER request frequency offset. F floatENTER set frequency offset, range +383 to -383 Hz, do not save in EEPROM. S floatENTER set frequency offset, range +383 to -383 Hz, save in EEPROM. f sending 2d 04 00 29 reported offset=0.000000 reported_frequency=10000000.0000000000

10,000,000.0000000000 Hz

Oh well....

Reply to
Jan Panteltje

Rubbitin frequency standard arrived and running:

This little interface box has the 5V switcher-regulator, the lock LED, the RS232 interface connector, and the green wire is the 10MHz output.

formatting link

Switcher is based on LM2596-33 with series resistors to get to 4.99V output. Black box stays cold, LED is via 3k3 from 'lock' output of the FE5680 to +5, so 'on' if locked.

The Toshiba 15V ebay supply delivers the power to this box. Need to think of a decent connector for that 10MHz... Been on now for more than an hour, see if temperature stabilized. Yes, that D connector was so neat soldered to the wires I left that in...

Move on...

Reply to
Jan Panteltje

module

crystal

mode

sex)

assumptions,

limiting.

many cases,

would melt or deform.

OK, it's useless to have a discussion with someone who has all the answers. If you're happy with your solution, I shouldn't intrude. I'll do my best to stop trying to help you. Carry on!

Reply to
mike

On a sunny day (Sun, 02 Dec 2012 08:05:06 -0800) it happened mike wrote in :

module

crystal

mode

sex)

assumptions,

limiting.

many cases,

would melt or deform.

Still no numbers, DID YOU READ THE SPECS? All is working well here BTW. See other postings.

Reply to
Jan Panteltje

Rubbitin frequency standard arrived and running:

the next step was to calibrate my little frequency counter:

formatting link

Now it shows about 9,999,999 Hz at the 10 MHz from the Rubbitin unit. It was 270 Hz low.. (showed 10,000,276 Hz). Soldered small SMD caps to get the crystal frequency just right, have to wait every-time after soldering as these caps are temperature dependent. At room temperature is should now be OK, 'calibrated'.

As I use it with a :256 prescaler this gives me 256 Hz error on 2.56 Ghz. Good enough for 'in band' check on both GPS and whatever. Useful little counter to have.

Rubidium unit on all day, works perfectly.

Reply to
Jan Panteltje

Rubbitin frequency standard arrived and running:

As all is working and as I now learned how to control it, it was time to take the thing apart, take some pictures, useful for finding stuff in the circuit (looking for the 50.255055 MHz.).

Please, these are very high resolution images of the PCBs, some more than 4 MB, so if you are on a low bandwidth connection do not click these links: Main board with AD9832 synthesizer:

formatting link
connector J8 also seems to carry 10MHz or there about, Higher frequency can be found next to the AD9832 on one of the vias... Now that is strange, IIRC Dave's unit had an AD9830??? This chip has a max clock input frequency of 25 MHz, maybe they did a divide by 2 somewhere... Still have to find that 50.255 on this board.

Back side board:

formatting link

Physics board, perhaps this is preamp and heater control?

formatting link

Put it all together, it is interesting, while it is 'locking', so not locked yet, the frequency sweeps over quite a wide range (relative to 10.0 MHz), it flies past the right value a couple of times (up down), and then finally locks in Still working :-)

It is fun to have this very precise reference around.

Reply to
Jan Panteltje

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.