GPSDO project

[throbscottle]

Ahh - I love a short answer!

 

[throbscottle]

OCXO arrived today. It is still attached to the board it was used on originally. Going to wait till the weekend so I have time to remove it carefully before testing!

 

[dr pepper]

Goody.
Looks like 2008 vintage, thats not necessarily a bad thing, oscillator aging might have setlled down by now.
Hot air paint stripper will get it off, or dremel the board into sections & solder iron each pin off.
Mine had a label with a screw underneath so I could tune the control voltage mid way for correct freq.

 

[throbscottle]

Yeah I read the post-y2k ones are better. This model doesn't have an adjustment screw. I'm a bit wary of using the solder-sucker because of the vibration.

 

[dr pepper]

I just got a voltage controlled 10Mc tcxo from Rs components, less than 4 quid and 2ppm accuracy over its temp range.
The pins on your ocxo will take a smack from a solder sucker, they are usually moulded into a plastic substrate, fatigue from bending is the only thing you need to be concerned with.

 

[throbscottle]

My GPS module arrived yesterday. Tiny! The whole eBay module is the size I expected the Neo module alone to be!
Just a few bits to order from RS now - I've been holding out in case of basket additions, save packaging....
The oscillator came off the board really easily. Bit of rust underneath so it's got wet at some time. I was concerned about shock because when I was looking into them, some specify an impact force and it's quite low - I suppose because jarring the crystal might affect its accuracy. Anyway I put it on a few layers of bubble wrap and used the sucker anyway. Found my guessed footprint was about 30% too big. Warms up at nearly 2A and settles down to about 230mA. Nice looking sine out about 3V, I was expecting much lower level about 200-300mV. I did check the reference voltage but can't remember now what it is - might be as high as 5v.

 

[dr pepper]

Crystals can break, I've changed a few broken ones, takes a good clout though, you'll know more when you have it locked to gps, if the control voltage is 2.5v or around its fine, and I'd say it most likely is.

My first neo went pop, the second one is only as big as the antenna square thingy.

The piezo ocxo only pulls a couple of hundred mA, but its a 24v heater, the fact the current drops shows its probably working fine, the case should get warm after a while too, they are meant to leak a little heat so the controller can regulate.

 

[throbscottle]

I've just been trying the Neo on my laptop. Took a while due to running U-Center on Wine which will no longer create the correct com port - and it deletes them when created manually if you do it before loading the application. I think I may need to file a bug report...
Anyway, it got a gps fix and I was able to have a play with it. So it works anyway. Now I just need to figure which of all these dozens of options are relevant (apart from the timepulse).

 

[throbscottle]

Dammit dammit dammit dammit dammit dammit!
I wanted to see if the eBay module with the Neo7 on it needs 3v to run or if it has an on-board regulator so it can run on 5v. So I started with a low voltage on the module's Vcc pin and turned it up whilst testing the voltage on the Neo7's Vcc pin. Stopped at 3.3v. So I don't need a separate 3v regulator. But then when I connected it to the laptop so I could set the PPS signal again and test it, U-Center couldn't connect to it. Seems I did something it didn't like and now it's dead (with the PPS light having gone from being always on to always off)
I'm just hoping it's a support component and not the Neo7 itself. 25Ω between Vcc (the 3.3v one) and ground doesn't look good.

 

[dr pepper]

Oops.
I blew one of my neo 6's, cant remember what I did now, possibly got nuked by 12v.
I have killed a couple of usb to 3.3v rs232 converters by stuffing 5v rs232 up them, that mighta killed your neo, and the serial comes from the central chip so it would be game over, allthough the led would probably still flash so it might not be that.
25 ohm across the rails isnt necessarily a problem, check the current draw, should be 100ma or so.

 

[throbscottle]

Thing is I had nothing connected, just power. I did notice a kind of smokey wisp going up but afaict the worst I did was short the supply or possibly adjacent pins but nothing bad should result. I checked the regulator afterwards and it seems fine. I'm hoping I'll find something failed between the reg and the neo though there's not much if anything.

 

[throbscottle]

Updated design. I got most of the parts from RS now but forgot the LM350T. I took out the 3v regulator since I've established the module will run on 5v.
Don't know how much difference the ambient temp sensor will make, I haven't got round to testing the oscillator's current at different ambient temps yet.
The loop filter is still very much a placeholder. I'm getting there slowly...
The memory backup battery should be good for a year. It's at the lower end of it's trickle charge range, or at least it should be.

 

[dr pepper]

Thats looking more like it, I like the lock detector.
One of mine has a cheap chinese digital voltmeter on the ocxo voltage, its very handy as I know when its at 2.50V everything is working.
My oven also has a warmed up output, my system has startup switches which speed up the loopfilters to make them lock, if I'dve thought about it a bit better I'dve used the warmed up signal to drive a 4066 or reed relay and made this automatic.
My new gpsdo whould arrive in the post tomoz.

 

[throbscottle]

Ha ha, I copied the lock detector from the 4076 which has it in place of PC3
Those cheap Chinese digital voltmeters chuck out a lot of noise. But maybe it has no effect how it's connected.
Yes a warmed-up output would have been good. I'll just have to see how this arrangement goes. The Piezo oscillators have them, but for my budget I read that these Trimble ones were better.

I'm feeling especially stupid over blowing my module. I realised later all I needed to to was see if 5v or 3.3v appears on the Vcc pin when it's connected to USB. IMA Dummass...

In my latest project I'm thinking of using a 'hc4046, however I need to be able to force phase comp 2 o/p to go high impedance regardless of its inputs, messing with the inputs is the only way I can think of doing it.

Can't you just feed it through a tri-state buffer?

 

[dr pepper]

I looked up 4076, google tells me its a quad d type flip flop.
I have various power rails in my setup, the meter is on the o/p driver chip supply, but I'll check for that when I get home tomoz.
One of mine is a piezo, another handy thing it has is a screwcap with a trimmer behind so you can set the loopvoltage bang on 2.5v when its in phase, a good nicety.
You know what they say, poo happens just get another.
The 4046, 'hc4046 has a tri state buffer on the Pc2 output, thats part of how it works, I'm going to test the thing on breadboard, I think with the chips enable input inactive it will be in high impedance, a tri state buffer is another good way as it regenerates the logic level, there is on the board some spare ways in a 4066 so I could use those.

 

[throbscottle]

Sorry I meant 74hc7046. Duh...
I knew the hc4046 has a tri-state buffer on PC2, I only mentioned using an external one since you mentioned wanting to control it separately.

 

[dr pepper]

Well you told me soething there, I knew there was another version of the 4046 but didnt know it was the 7046, at first I thought the lock detect was a clever IQ type detector like in the Ne567, but its just a simple retiggerable timer, still it'll save a few extra parts.

Looking at the datasheet and laterally thinking theres another way to force Pc2 o/p to high impedance, just force both phase inputs to the same logic level and make it think the 2 signals are in phase and set the o/p high impedance.

 

[throbscottle]

There's also a 74HC9046.
The 7046 and 9046 are both quite expensive though.
Sounds like a good idea you have there.

 

[throbscottle]

It occurred to me I could use a window comparator to see if the oscillator is approaching EOL - it the control voltage gets close to ground or vref it is reaching it's limit of adjustment and should be replaced. That's what the time-nuts folks reckon anyway.
I doubt it would ever get there with my usage - more a case of completeness than need...

 

[dr pepper]

Thats why I have the meter on the control voltage, if it changes you can tell straight away, and you can tell when its warmed up.
Switching lights on affects the loop voltage, because my sync signal is 60kc.
I'm working on an arduino monitoring system, to decode the time and measure / compare frequencies, as well as display loop voltages etc.
Might build a system that monitors all 3 systems, Msf, Gps and Radio4.