Siglent SDG1000X Series Clock Adjust

[Wrenches of Death]

This question concerns an SDG1032X that I purchased last year.

The internal 10Mhz clock is off by about 130hz. It's a little annoying, especially if using the counter function,
but is probably within spec. I generally use a much more accurate external time base when appropiate, so I'm not going to lose any sleep over it.

My question is, is it possible to zero the internal time base to an accurate external timebase with keyboard commands?

It seems that this would be a handy feature. If it isn't currently available, perhaps it can be included in a future firmware update.

Thanks in advance!
WoD

[Tjuurko]

Now there is an ordinary resonator for 25 MHz. Do not hope to remotely adjust frequency.

[Wrenches of Death]

I see no traditional adjustments anywhere on the main board. That would lead me to believe that the calibration of this instrument is handled "closed box" via either the network or USB port.

If that is the case, then it might be possible for keypad access for something as simple as the clock adjustment.

My fingers are still crossed! 

WoD

[Domitronic]

Have a look at the service manual:

https://www.batronix.com/files/Siglent/Funktionsgeneratoren/SDG1000X/SDG1000X_ServiceManual_EN.pdf

At first glance it doesn't look like frequency can be adjusted.

[Wrenches of Death]

Very interesting, thanks!

Looking through the information in the Python calibration scripts, it seems that the clock is only adjustable on the SDG2000 and up. Lots of other adjustments though.

It's still disappointing that I'll have to live with the 130hz error until the warranty has expired.

Thanks again!

WoD

[Hexley]

You raise an interesting question about software correction of the clock. Of course that is not possible in the sense of having a DAC tweak the control voltage on the internal crystal oscillator, because the SDG1000X does not have that kind of hardware.

However, even with the existing hardware it might be possible to adjust the parameters sent to the DDS engine in order to tweak the output frequency. So if, for example, the internal oscillator is known to be100 ppm slow then the firmware could change the DDS arguments so that it operates 100 ppm faster than nominal. Then the offsets should cancel. That would have the same result as actually adjusting the crystal frequency.

The offset data could be entered manually if one had measured the AWG's output against a precision frequency reference and noted the error, as you did. It might even be possible for the instrument to use its internal counter to measure a precision input and do the necessary calculations to learn the offset of its internal reference.

So it might be possible in theory. But probably not at the top of the list of things that the firmware team should address in the next release.

Maybe a more practical solution is to buy a $10 OCXO from eBay, build a little board that runs from a 5V USB charger, and use that setup to drive the rear panel reference input. One afternoon's work and you should be good to better than 1 Hz at 10 MHz. Maybe much better; depends on the OCXO.

[Wrenches of Death]

Maybe a more practical solution is to buy a $10 OCXO from eBay, build a little board that runs from a 5V USB charger, and use that setup to drive the rear panel reference input. One afternoon's work and you should be good to better than 1 Hz at 10 MHz. Maybe much better; depends on the OCXO.

That's very clever and would work. I've already got a half a dozen or so things I can feed it off of. It makes it a lot less portable though.

When the warranty expires, if I'm still kicking, I'll go ahead and install a better oscillator in the thing. The board already has the holes for one. You'd think that for what this thing cost, it would have had something better than a 50ppm clock in it, or at the very least, a way to adjust it. Oh well, live and learn I guess.

The lack of service data on current test equipment is the primary reason I try to buy as little of it as possible.

Even Simpson quit supplying schematics and a parts list on their 260 multimeter. It's not like it contains any "top secret" proprietary circuitry or anything. It's an analog multimeter. Anybody with more than a handful of brain cells to rub together and a calculator can work out the values of every component in it if they want to. It's not rocket science.

Thanks!

WoD

[Ringmodulator]

AFAIk, there is a footprint for a tcxo on the mainboard f the SDG1000X series.
There is a foto in the SDG1000x main thread, but I have not seen confirmation, that someone actually did the mod.

I did the modification to a SDG1025. The tcxo has a trimcap, so it can be finetuned.

Chris

[Wrenches of Death]

Thanks! I must have seen the same picture. I'm hesitant to cut the seal on mine and kill the warranty though.

I've been thinking about the post to this thread by Hexley about the external clock and it got me to thinking. A small board with an external oscillator that is powered not from a charger, but from the USB port on the back of  the SDG10032X would be the way to go. Two short cables and some Velcro and portability is accuracy and portability are back. You could even move the oscillator to another instrument if you want.

Would you have the part number handy of the oscillator you used in the SDG1025? I'd like the get the dimensions and see if I can scale the footprint in the picture of the SDG1032X board.

Thanks again!
WoD

[2N3055]

I think smart thing is not to open it. It has 10MHz ref in.
If best accuracy is needed, maybe putting in an effort to get GPSDO. That would also serve a reference clock source for other purposes.

[Wrenches of Death]

I think smart thing is not to open it. It has 10MHz ref in.
If best accuracy is needed, maybe putting in an effort to get GPSDO. That would also serve a reference clock source for other purposes.

I only use the Siglent for it's convenience and portability.

I've got a GPSDO that feeds my 5345A. It's a pretty decent combo for accuracy but kind of clunky to lug around. I've had the 5345A for about twenty five years. I replaced all of the filter caps in around 2000 or so. It might be a good idea for me to do that again. I've had no other problems with it.

I also have a 5360A that I've had must be close to thirty years now. It's an even bigger pain to lug around. With the plugin, it's got to be at least sixty pounds. The time base in it is pretty decent too, but only 5Mhz. I think it will recognize and run with a 10Mhz external one but have never tried it. I noticed the last time I used it, it took noticeably longer than usual for the nixies to fire. I don't know if the neon originally had an isotope to help ionization that has had a half a dozen half-lives, or if the power supply is going to need work.  Yet one more project...

The SDG1032X counter function is convenient and will display simple statistical info without the hassle of a computer but it has to be used off of something other then the internal clock. It's running right now off of the GPSDO monitoring the output of an old Bulova PCOXO oscillator that I used to use as my standard around here for a couple of decades. Make a tweak, clear the readings and let it run a day. Rinse and repeat. It keeps me off of the streets.

Thanks!

WoD