Siglent SDG1025 Arbitrary/function generator under some tests

[mimmus78]

After a fast check with Siglent, firmware file SDG1000-V100R001B01D01P27.ADS was good also for updating my:

 - SDG1025
 - SOFT.V 1.01.01.23
 - HW.V 02-00-00-19-24[/li][/list]

to the latest version (27).

[RJdaMoD]

Hello,
i'm new to eevblog forum and was looking for an inexpensive arb-gen to fulfill my hobby-needs. Thanks to  aghp and his extensive testing i consider buying an SDG10xx.
I'm unsure about one thing though: The specs state that arbitrary waves can be generated with frequency up to 5MHz. Does this mean,
a) that the actual bandwidth in arb-mode is 5MHz, that is, a 5Mhz lowpass is switched between dac and bnc?
b) that the repetition rate of the arb-pattern is 5MHz at maximum, but the signal itself can have higher bandwidth?
In the last case, what is the maximum bandwidth in arb-mode?
At 125MSa/s, i assume that the analog circuitry will  diminish the theoretical 62.5MHz, but i would be really interested how far. Does the stated noise bandwidth relate to this?
From aghp`s tests a saw a rise time of 8ns for pulses with the SDG1010, which would equate to about 45MHz, while noise should be up to 10MHz.
I'm now somehow curious to know what happens in arb-mode. Could someone test this? I have attached a csv with 16 square cycles which would give a 62.5MHz square at 3.90625MHz, which of course should give a sine at this frequency, but lower would be interesting for me to see.
Thanks in advance!
Greetings from snowy germany ,
Roger

[rf-loop]

Just write Sin(10*x)x  in equation window and move this waveform .csv (10 cycle of sine) to generator arb memory using USB stick or USB cable. Use it and select 5MHz freq. You get 50MHz out. (attenuated). (vertical scaling is automatic to full p-p in memory)

Risetime is <8ns

Somewhere I have frequency response (level) tests but this time I can not find these and other works keep me busy now.

So in your text

b)  is right.

[RJdaMoD]

Thanks!
I currently have no access to an SDG, but i want to buy one and wonder if i should buy the 10 or 25MHz version. For my hobby needs,10 would be sufficient, but i would like to have the possibility to go higher.
So if arb does the job, it will be the 1010.
But what is the difference to 1025 then? Just firmware, or better parts in the signal path?

PS: I tried equation draw in easywave (1.7), but somehow it didn't work yesterday. Strangely your example works, although i could swear i tried the same yesterday, and it didn't work. Thus i generated the csv with another program, but essentially it is sin(16*x)/abs(sin(16*x)).

[RJdaMoD]

Hi, me again, my 1010 just arrived. It works pretty well, came with fw23. Since hw version is that of mimmus78, i could successfully apply the posted fw27 (thanks btw!). Arb bandwidth is ~62.5MHz as expected (of course just sines at this frequency), but there is some jitter in that range. Seems as if the dac clock is not that stable, but it's enough for me.
I have one question though: I can load arb waves via usb stick, and the SDG outputs the loaded wave as it should, but it is not saved in the selected arb place. Only way to store waves in the arb memory seems to be via easyWave. Can someone confirm this? Seems like a bug to me.
Greets, RJ

[Siglent]

Thank for you attention.
The question that you doubt is normal.
"load arb waves via usb stick" will store the "stored wforms" that distinguishes the arb wave store

[felixd]

Hello. I'm looking for some information about max Vpp of 10MHz ext signal. I tried to find in in all PDF files provided for SDG1025 but failed.
What I found is only max voltage information of AM ext modulation.

If my ref 10MHz signal is at 1.5 Vtop level should I convert it to TTL/CMOS level for Siglent?

For an external source, the depth of AM is controlled by the voltage level of
the connector connected to the [Modulation In]. ±6V corresponds to the
currently set depth 100%.

2nd question related: Is there any factory voltage standard for 10MHz reference signal?

Edit: http://digital.ni.com/public.nsf/allkb/FA7810B075EF6A9F86256E5300616E51

[tautech]

Sorry I also could not find a definitive answer in my documentation.
There is mention of +/-6V in Chapter 2.7 of the user manual but this mention applied to use of modulation and this is probably just what you found also.
It should be safe to use these levels as max until we get confirmation from Siglent after the weekend.

But I'd suspect with only 1.5Vtop you'll need to use some level shifting to reach the TTL levels specified as needed for external inputs.
http://www.interfacebus.com/voltage_threshold.html

More info can be found in this thread:
https://www.eevblog.com/forum/testgear/the-sdg1000-and-sdg800-thread/

[kwass]

Hello. I'm looking for some information about max Vpp of 10MHz ext signal. I tried to find in in all PDF files provided for SDG1025 but failed.
What I found is only max voltage information of AM ext modulation.

If my ref 10MHz signal is at 1.5 Vtop level should I convert it to TTL/CMOS level for Siglent?

For an external source, the depth of AM is controlled by the voltage level of
the connector connected to the [Modulation In]. ±6V corresponds to the
currently set depth 100%.

2nd question related: Is there any factory voltage standard for 10MHz reference signal?

Edit: http://digital.ni.com/public.nsf/allkb/FA7810B075EF6A9F86256E5300616E51

This thread discusses the same issue: https://www.eevblog.com/forum/testgear/using-the-racal-dana-1992-as-an-external-reference-for-the-siglent-sdg1025/


If you need higher clock accuracy without using an external 10MHz source, you can pretty easily modify the SDG1025 to use an internal 25MHz TCXO.  I've done this and it works great for most stuff so that I only occasionally need to use the 10MHz input.  Here's how to do it: https://youtu.be/RxCDfNCLj08

[felixd]

@tautech: Thank you for your link.
@kwass: Thank you for your link as well. I'll use than 74HC14 to shift it to 3.3V.

[Siglent]

Hello. I'm looking for some information about max Vpp of 10MHz ext signal. I tried to find in in all PDF files provided for SDG1025 but failed.
What I found is only max voltage information of AM ext modulation.

If my ref 10MHz signal is at 1.5 Vtop level should I convert it to TTL/CMOS level for Siglent?

For an external source, the depth of AM is controlled by the voltage level of
the connector connected to the [Modulation In]. ±6V corresponds to the
currently set depth 100%.

2nd question related: Is there any factory voltage standard for 10MHz reference signal?

Edit: http://digital.ni.com/public.nsf/allkb/FA7810B075EF6A9F86256E5300616E51

Hello Felixd,
    There is the data about max Voltage of 10MHz External trigger input signal:(For SDG1000)

    Refence Frequency Input   
Voltage level           5Vpp--5.5Vpp
Frequency range   10MHz±1kHz
Input impedance   ?5k?.AC coupling

Hope it can help you.
Best regards.

[felixd]

Of course it helps. Thank you so much

    There is the data about max Voltage of 10MHz External trigger input signal:(For SDG1000)

    Refence Frequency Input   
Voltage level           5Vpp--5.5Vpp
Frequency range   10MHz±1kHz
Input impedance   ?5k?.AC coupling

[smbaker]

If you need higher clock accuracy without using an external 10MHz source, you can pretty easily modify the SDG1025 to use an internal 25MHz TCXO. 

Silly noob question -- why is he preheating the PCB before soldering and desoldering? Is it due to the precision of the components?

[nctnico]

It just makes it easier to solder components in a multi-layer board especially if you have a low power iron. If you have a 80W iron it shouldn't be a problem unless the thermal reliefs are absent.

[kwass]

It just makes it easier to solder components in a multi-layer board especially if you have a low power iron. If you have a 80W iron it shouldn't be a problem unless the thermal reliefs are absent.

I had no problems at all with a 40 Watt station and no preheating.

[smbaker]

It just makes it easier to solder components in a multi-layer board especially if you have a low power iron. If you have a 80W iron it shouldn't be a problem unless the thermal reliefs are absent.

Ah, I see. I have a hot air station, so maybe I'll do that.

Now the only questions are whether or not to bother with it (I suppose more precision is always nice) and where to get the TCXO. The ebay listing mentioned in the video isn't there anymore. EBay has lots of 3.3V TCXOs, but a scarcity of 5V ones.

[tautech]

It just makes it easier to solder components in a multi-layer board especially if you have a low power iron. If you have a 80W iron it shouldn't be a problem unless the thermal reliefs are absent.

Ah, I see. I have a hot air station, so maybe I'll do that.

Now the only questions are whether or not to bother with it (I suppose more precision is always nice) and where to get the TCXO. The ebay listing mentioned in the video isn't there anymore. EBay has lots of 3.3V TCXOs, but a scarcity of 5V ones.

Here's some info for you:
https://www.eevblog.com/forum/testgear/the-sdg1000-and-sdg800-thread/msg614956/#msg614956

[kwass]

Now the only questions are whether or not to bother with it (I suppose more precision is always nice) and where to get the TCXO. The ebay listing mentioned in the video isn't there anymore. EBay has lots of 3.3V TCXOs, but a scarcity of 5V ones.

I'll bet that this one will work fine:  http://www.ebay.com/itm/1pcs-TCXO-0-1ppm-25-000MHz-25MHZ-Ultra-precision-Gold-Oscillator-FOR-audio-DIY-/131649190505?hash=item1ea6e73a69:g:uFgAAOSwQItUF9Lw

[smbaker]

I'll bet that this one will work fine:  http://www.ebay.com/itm/1pcs-TCXO-0-1ppm-25-000MHz-25MHZ-Ultra-precision-Gold-Oscillator-FOR-audio-DIY-/131649190505?hash=item1ea6e73a69:g:uFgAAOSwQItUF9Lw

I saw that one. I wish it had a part number and a datasheet that would lend some legitimacy to it. The most I could find was a thread on another forum speculating that it's another manufacturer's part that's been gold-plated and remarked, and sold at a premium to unsuspecting audiophiles (some of whom will buy anything so long as it's gold). I'm not sure if that's true or not, but this part makes me suspicious.

[kwass]

I saw that one. I wish it had a part number and a datasheet that would lend some legitimacy to it. The most I could find was a thread on another forum speculating that it's another manufacturer's part that's been gold-plated and remarked, and sold at a premium to unsuspecting audiophiles (some of whom will buy anything so long as it's gold). I'm not sure if that's true or not, but this part makes me suspicious.

How about from here:  http://www.aliexpress.com/item/ONE-PIECE-PRECISION-0-3ppm-25MHz-25-000MHz-25M-DIP-TCXO-Support-bulk-orders/32297054283.html . 
This looks identical to the one I bought (says M.D.D on it)

[smbaker]

How about from here:  http://www.aliexpress.com/item/ONE-PIECE-PRECISION-0-3ppm-25MHz-25-000MHz-25M-DIP-TCXO-Support-bulk-orders/32297054283.html . 
This looks identical to the one I bought (says M.D.D on it)

Thanks, I went ahead and ordered a pair (always like to have an extra for something that takes so long in shipping) of those. When they arrive in a month or so, I'll perform the upgrade.

Scott

[electrongeek]

I am perhaps going to show a bit of noob ignorance here, but reading through the SDG1000 service manual, I don't see where a frequency calibration is performed. Or rather how the unit might be calibrated by a high accuracy/precision 10 MHz source, like my GPSDO. Can someone clue me in? Thanks.

[tautech]

I am perhaps going to show a bit of noob ignorance here, but reading through the SDG1000 service manual, I don't see where a frequency calibration is performed. Or rather how the unit might be calibrated by a high accuracy/precision 10 MHz source, like my GPSDO. Can someone clue me in? Thanks.

Welcome to the forum.

I imagine there will be parameters on Self cal that it need meet to pass.
To switch between internal clock and external use the ClKSource function in the Utility menu.

[rf-loop]

I am perhaps going to show a bit of noob ignorance here, but reading through the SDG1000 service manual, I don't see where a frequency calibration is performed. Or rather how the unit might be calibrated by a high accuracy/precision 10 MHz source, like my GPSDO. Can someone clue me in? Thanks.

Afaik, there is not  adjustment for internal frequency reference oscillator.

Accuracy Within 90daysv±50 ppm within 1 year ±100 ppm 18°C ~ 28°C
Temperature coefficient <5 ppm/°C

As long as it is inside specs it do what is promised and ok. If main board TP4 is out from specified 25MHz accuracy, then you write over equipment "Failed, need repair".

If want more accuracy there can use external reference.

(It can also modify using much higher accuracy TCXO or OCXO or DOCXO. Most easy mod is if use 25MHz models)

[electrongeek]

Wonder what the factory does to cal the 25 MHz clock, or do they make the assumption it is "close enough" from the start?

Just surprises me a little bit.

Thanks for your help gang.