Author Topic: The Siglent SDG2042X Thread  (Read 275244 times)

0 Members and 1 Guest are viewing this topic.

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16025
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #375 on: May 17, 2016, 08:20:15 pm »
Yeah, just see the blatant ripoff of the Agilent multimeter copied by Siglent :P
You'd be well advised to do some more homework.  :P

See this post and others that are linked to.  ;)
https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/msg637615/#msg637615
Avid Rabid Hobbyist
 

Offline Aeternam

  • Supporter
  • ****
  • Posts: 97
  • Country: lu
    • Schartz Engineering
Re: The Siglent SDG2042X Thread
« Reply #376 on: May 25, 2016, 07:04:18 am »
I had just opened up my new SDG2042X to see if it would be possible to give it isolated outputs (with a little add-on board containing the data isolators and a DC/DC converter).

Any news? This seems like an interesting little project...
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16025
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #377 on: May 25, 2016, 07:57:05 am »
I had just opened up my new SDG2042X to see if it would be possible to give it isolated outputs (with a little add-on board containing the data isolators and a DC/DC converter).

Any news? This seems like an interesting little project...
Are you aware the SDG5000 series offers isolated outputs as standard spec?
Avid Rabid Hobbyist
 

Offline Timpert

  • Contributor
  • Posts: 43
  • Country: nl
Re: The Siglent SDG2042X Thread
« Reply #378 on: May 25, 2016, 11:47:17 am »
Quote
Are you aware the SDG5000 series offers isolated outputs as standard spec?
Yes, I am, but the SDG5000 is in another price bracket, and it has a slightly poorer sine wave quality too. So getting an SDG2042X and trying to isolate its outputs seemed like a worthwhile attempt to me.

I will share some news on my progress, but please don't expect too much on short notice. I have gotten quite a bit of work on my plate lately.
 
The following users thanked this post: rf-loop, tautech

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3069
  • Country: fi
  • Starting with DLL21
Re: The Siglent SDG2042X Thread
« Reply #379 on: May 25, 2016, 03:01:47 pm »
I had just opened up my new SDG2042X to see if it would be possible to give it isolated outputs (with a little add-on board containing the data isolators and a DC/DC converter).

Any news? This seems like an interesting little project...
Are you aware the SDG5000 series offers isolated outputs as standard spec?

They are not true isolated if mean true DC-AC-(RF) isolation.
But, DC isolation, they do not lie. But....

Also channels are not isolated from each other.

In this case "Isolation" mean that whole signal board is DC floating but nothing more if look floating impedance.

There is high capasitive connection to chassis and between chassis and BNC GNG is max 42V.
Capasitance from BNC GND to Chassis is somewhere around >3uF ! (least in SDG5082 individual unit what I have)
Reactance with 50Hz is around 1ohm Omg. Err...  Right is of course around 1kohm.


Example HP33120A  floating output have around 4000pF  as can see in attachment.
This reactance with 50Hz is around 800kohm
Same max 42Vpk.

Lets hope Timbert find nice solution and sell it to Siglent.

« Last Edit: May 28, 2016, 07:19:52 am by rf-loop »
If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 

Online MasterTech

  • Frequent Contributor
  • **
  • Posts: 888
  • Country: 00
Re: The Siglent SDG2042X Thread
« Reply #380 on: May 25, 2016, 04:47:08 pm »
One thing i'm missing from the SDG2042 is the ability to add noise to the generated waveform, ie sine+noise, pulse+noise....

BTW heres the spectrum of the white noise coming out of the generator, 130mV stddev over 50ohms, 120Mhz bandwidth  :D:
 

Offline Timpert

  • Contributor
  • Posts: 43
  • Country: nl
Re: The Siglent SDG2042X Thread
« Reply #381 on: May 27, 2016, 08:03:53 am »
Quote
Reactance with 50Hz is around 1ohm
Really? How can that be with a 3 uF capacitance to ground? And where is the component that causes that? Anything like that is borderline useless, as one of the points of isolated outputs is to break the ground loop that now always exists when using the generator together with a scope. I have looked at the teardown video and pictures of the SDG5000, and they use a bunch of ADuM isolators, and galvanically isolated power supplies, so a reasonable capacity to ground would be somewhere around 10 nF.

I have opened up my generator a while ago to figure out the signals between the CPU board and the signal generator board. I am currently still doing the feasibility research, but so far I have not found anything that makes isolation impossible. It is however a little less trivial than I initially thought. When I have established definitively to go ahead with this, I'll open up another thread for it.

MasterTech's post is one of many showing that the signal quality of this generator is very good for it's class.
 

Offline nugglix

  • Regular Contributor
  • *
  • Posts: 209
  • Country: de
Re: The Siglent SDG2042X Thread
« Reply #382 on: May 27, 2016, 08:51:26 am »
One thing i'm missing from the SDG2042 is the ability to add noise to the generated waveform, ie sine+noise, pulse+noise....

*cough*

Under modulation you can find noise as mod-waveform.
At least on my unit...
Works for AM, FM and PM at least.
Also seems to work for the standard waveforms.
But I've no time to test that now.
 

Offline Ivan7enych

  • Regular Contributor
  • *
  • Posts: 136
  • Country: ru
    • My astronomy projects
Re: The Siglent SDG2042X Thread
« Reply #383 on: May 27, 2016, 10:52:57 am »
One thing i'm missing from the SDG2042 is the ability to add noise to the generated waveform, ie sine+noise, pulse+noise....

BTW heres the spectrum of the white noise coming out of the generator, 130mV stddev over 50ohms, 120Mhz bandwidth  :D:
Spectrum looks like an analog based noise generator.

Rigol DG4062 can add noise to signal, but it's spectrum is not uniform and looks like many many spikes (digitaly constructed noise pattern?).
 

Online MasterTech

  • Frequent Contributor
  • **
  • Posts: 888
  • Country: 00
Re: The Siglent SDG2042X Thread
« Reply #384 on: May 27, 2016, 12:27:18 pm »
The lowest signal tone it can generate is 1 mVpp, -56dbm and its pretty clean:


 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3069
  • Country: fi
  • Starting with DLL21
Re: The Siglent SDG2042X Thread
« Reply #385 on: May 27, 2016, 03:05:56 pm »
Quote
Reactance with 50Hz is around 1ohm
Really? How can that be with a 3 uF capacitance to ground? And where is the component that causes that?

Here attached image from one SDG5082 output on the PCB.

Capacitor parallel with bidirectional TVS  from connector GND to chassis.
If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 

Offline Timpert

  • Contributor
  • Posts: 43
  • Country: nl
Re: The Siglent SDG2042X Thread
« Reply #386 on: May 27, 2016, 09:28:37 pm »
Have you measured the AC voltage between signal ground and chassis on your SDG5000 with it powered on, output enabled and nothing connected? Also measure the AC leakage current from signal ground to chassis. I bet there's some capacitive coupling from the mains to the output in the power supply via the transformer and the EMI suppression cap. Or your TVS diode may be shot.
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3069
  • Country: fi
  • Starting with DLL21
Re: The Siglent SDG2042X Thread
« Reply #387 on: May 28, 2016, 07:15:01 am »
I bet there's some capacitive coupling from the mains to the output in the power supply via the transformer and the EMI suppression cap. Or your TVS diode may be shot.

TVS diodes are ok. I told there is high capasitive coupling from signal board to equipment chassis. These coupling capasitors, parallel with TVS, you can see in image. I do not need advisory for find problem in my SDG5k. Why start finding problem where is not problem?  I only tell how it is designed (and when asked about this capasitance) and it works as designed and manufactured.  I'm not interested to find where is possible other capasittive pico- and nano Farades connections (designed and parasitic)  here and there when there exist also microfarads.)  Why you think TVS diode shot, my SDG5k does not have a fault.  Or do you mean this 1kohm capasitive reactance between chassis and output BNC GND. 1kohm 50Hz reactance is not at all good.  It is barely acceptable if we have pure clean sinewave mains.  There exist lot of LF-HF noise in mains and with this high capasitive connection if there exist (example) 50kHz freq components reactance is now around 1ohm (take scope or analyzer and look what all there is traveling in mains (this is why I use in many cases filtered mains). When look example HP in previous example, there is around 4nF.
So in SDG5k we can talk more like there is just "DC block" <42V

Totally other question is if I like it or not or if someone else like it or not.

I can tell my personal single opinion and it is: I do not like how they have done it. If I design this equipment I will do it different for bit better (in this specific output isolation question).

This "isolation" in SDG5k is still lot of better better than nothing.
« Last Edit: May 28, 2016, 07:21:30 am by rf-loop »
If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 

Online MasterTech

  • Frequent Contributor
  • **
  • Posts: 888
  • Country: 00
Re: The Siglent SDG2042X Thread
« Reply #388 on: May 28, 2016, 09:28:50 am »
Just found a SW bug regarding the modulation with noise:

1) Set carrier, eg 40MHz
2) turn on AM modulation, 100% depth, sine wave, set AM frequency to  x Mhz (I used x=1, but tested others too)
At this point the correct AM modulated waveform is generated
3) Change modulation "shape" to noise
instead of noise being modulated, an infinite number of sub-carriers spaced x Mhz apart are generated
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16025
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #389 on: May 28, 2016, 09:39:25 am »
Just found a SW bug regarding the modulation with noise:

1) Set carrier, eg 40MHz
2) turn on AM modulation, 100% depth, sine wave, set AM frequency to  x Mhz (I used x=1, but tested others too)
At this point the correct AM modulated waveform is generated
3) Change modulation "shape" to noise
instead of noise being modulated, an infinite number of sub-carriers spaced x Mhz apart are generated
Can I ask, which model? 40, 80 or 120 MHz unit.

Edit.
Got a PM from MasterTech and his is a improved SDG2042X. <wink>
We'll attempt to replicate this bug with a new unit and if proven will be notified to Siglent next week.
« Last Edit: May 31, 2016, 05:12:55 am by tautech »
Avid Rabid Hobbyist
 

Offline Timpert

  • Contributor
  • Posts: 43
  • Country: nl
Re: The Siglent SDG2042X Thread
« Reply #390 on: May 28, 2016, 02:13:19 pm »
@rf-loop: no offense meant, I just reacted to your post before you corrected your typo. 1k Ohms perfectly adds up with the 3 uF you mentioned. I hope we can agree that if the impedance to ground really were 1 Ohm, then something would be seriously knackered.

@everyone interested: The reason why Siglent chose the rather large capacitance between output ground and chassis for the SDG5000 lies in the EMI suppression scheme they chose for the power supply. There is a little 2.2 nF suppression cap next to the transformer, that is connected between the output common and the primary circuit's negative rail. This negative rail bounces up and down at roughly one half mains voltage (so in my case 325 V P-P), and if nothing is done to suppress that, that voltage will end up superimposed on the output voltage. Very high-Z though, but enough to give you a slight tingle and to damage electronics. When you draw that away with 3 uF to chassis it drops to some 80 mV RMS, low enough so that nobody will be bothered.

The SDG2000X power supply is almost identical to that of the SDG5000, and it has the same EMI suppression scheme. So I can choose to go for an isolation solution like the SDG5000 and accept 3 uF between output ground and chassis, or I can try to tackle the PSU as well, and aim for something in the 10 nF range, which a good bit more useful. But as rf-loop rightfully concluded, 3 uF output-to-ground capacity is still a lot better than no isolation. So I suppose this is one of the reasons why you pay top dollar for a Hagsight generator. 

 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3069
  • Country: fi
  • Starting with DLL21
Re: The Siglent SDG2042X Thread
« Reply #391 on: May 28, 2016, 05:21:55 pm »
@rf-loop: no offense meant, I just reacted to your post before you corrected your typo. 1k Ohms perfectly adds up with the 3 uF you mentioned. I hope we can agree that if the impedance to ground really were 1 Ohm, then something would be seriously knackered.

@everyone interested: The reason why Siglent chose the rather large capacitance between output ground and chassis for the SDG5000 lies in the EMI suppression scheme they chose for the power supply. There is a little 2.2 nF suppression cap next to the transformer, that is connected between the output common and the primary circuit's negative rail. This negative rail bounces up and down at roughly one half mains voltage (so in my case 325 V P-P), and if nothing is done to suppress that, that voltage will end up superimposed on the output voltage. Very high-Z though, but enough to give you a slight tingle and to damage electronics. When you draw that away with 3 uF to chassis it drops to some 80 mV RMS, low enough so that nobody will be bothered.

The SDG2000X power supply is almost identical to that of the SDG5000, and it has the same EMI suppression scheme. So I can choose to go for an isolation solution like the SDG5000 and accept 3 uF between output ground and chassis, or I can try to tackle the PSU as well, and aim for something in the 10 nF range, which a good bit more useful. But as rf-loop rightfully concluded, 3 uF output-to-ground capacity is still a lot better than no isolation. So I suppose this is one of the reasons why you pay top dollar for a Hagsight generator.

Yes this 1ohm was my typo.  Later I think why you wonder this. I read my msg agen and did not se any strange, until later I read agen and ...ooops..there is k missing.

But yes I agree what you say about construction (and also reason).  In equipment there need do compromise with things. If I do it I do different and try much less coupling cap to chassis as example there in HP.

But, with this design we need live with SDG5k.

It is nice if SDG2000X can get least some kind of isolation modification but if can reach 10-20nF range is "super" (in this case) imho.
If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 

Offline Timpert

  • Contributor
  • Posts: 43
  • Country: nl
Re: The Siglent SDG2042X Thread
« Reply #392 on: May 29, 2016, 07:52:52 pm »
For a moment I thought I might have a solution that brings isolated outputs with a capacity of roughly 60 nF to ground, with only a small change in the PSU's EMI filtering.  But some more thinking revealed an unacceptable side effect of this change: about 100 mVrms of residual at the switch frequency, ouch! So it will have to be 3 uF to ground and keeping the current PSU EMI filtering scheme, the same as with the SDG5000. Otherwise, EMI performance will be horrible.

The reason for this is the use of an unshielded power transformer in the (resonant) PSU, giving a significant capacitive coupling between primary and secondary. The deluxe option would be to replace the power transformer with an appropriately shielded one (2 shields are required at the least), but that would be a major operation. Not that I don't like the challenge, and I feel that it should be doable, but I also feel that I am going overboard when I take this route.
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16025
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #393 on: May 31, 2016, 05:19:57 am »
Just found a SW bug regarding the modulation with noise:

1) Set carrier, eg 40MHz
2) turn on AM modulation, 100% depth, sine wave, set AM frequency to  x Mhz (I used x=1, but tested others too)
At this point the correct AM modulated waveform is generated
3) Change modulation "shape" to noise
instead of noise being modulated, an infinite number of sub-carriers spaced x Mhz apart are generated
Can I ask, which model? 40, 80 or 120 MHz unit.

Edit.
Got a PM from MasterTech and his is a improved SDG2042X. <wink>
We'll attempt to replicate this bug with a new unit and if proven will be notified to Siglent next week.
Following on from the bug MasterTech revealed above....

Is a Noise type of modulation even required?  :-//
Why? Real use example please.

Do other AWG's provide Noise modulation of waveforms?
Makes? Models?

Avid Rabid Hobbyist
 

Online MasterTech

  • Frequent Contributor
  • **
  • Posts: 888
  • Country: 00
Re: The Siglent SDG2042X Thread
« Reply #394 on: May 31, 2016, 05:59:33 am »
Just found a SW bug regarding the modulation with noise:

1) Set carrier, eg 40MHz
2) turn on AM modulation, 100% depth, sine wave, set AM frequency to  x Mhz (I used x=1, but tested others too)
At this point the correct AM modulated waveform is generated
3) Change modulation "shape" to noise
instead of noise being modulated, an infinite number of sub-carriers spaced x Mhz apart are generated
Can I ask, which model? 40, 80 or 120 MHz unit.

Edit.
Got a PM from MasterTech and his is a improved SDG2042X. <wink>
We'll attempt to replicate this bug with a new unit and if proven will be notified to Siglent next week.
Following on from the bug MasterTech revealed above....

Is a Noise type of modulation even required?  :-//
Why? Real use example please.

Do other AWG's provide Noise modulation of waveforms?
Makes? Models?

It actually doesn't make any sense, if I were Siglent I'd remove that option in the modulation menus. They are looking for trouble by doing it this way.
However, adding noise to other waveforms (ie sine+noise, pulse+noise, etc...)  is very useful, as it lets you mimic real life conditions at the input of devices like logic gates, drivers, demodulators..... That feature should be added in my humble opinion. ^-^
 
The following users thanked this post: tautech

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3069
  • Country: fi
  • Starting with DLL21
Re: The Siglent SDG2042X Thread
« Reply #395 on: May 31, 2016, 06:22:18 am »
Just found a SW bug regarding the modulation with noise:

1) Set carrier, eg 40MHz
2) turn on AM modulation, 100% depth, sine wave, set AM frequency to  x Mhz (I used x=1, but tested others too)
At this point the correct AM modulated waveform is generated
3) Change modulation "shape" to noise
instead of noise being modulated, an infinite number of sub-carriers spaced x Mhz apart are generated
Can I ask, which model? 40, 80 or 120 MHz unit.

Edit.
Got a PM from MasterTech and his is a improved SDG2042X. <wink>
We'll attempt to replicate this bug with a new unit and if proven will be notified to Siglent next week.
Following on from the bug MasterTech revealed above....

Is a Noise type of modulation even required?  :-//
Why? Real use example please.

Do other AWG's provide Noise modulation of waveforms?
Makes? Models?

Take one SW radio and turn it on. Start searching some transmission.
What you hear. You hear enormous amount of natural and human made noise and somewhere inside this are also many kind of signals. In lab when we test example wireless data transmission we need signal + noise for emulate natural situation. (yes it can do also other ways) Some times overall noise level is far over wanted signal level and we need build equipments what can catch this signal from deep under noise level. ;) ;)

(Least I solve this problem with mixixing noise source and signal)

But overall, noise is very important tool for many many things and usefulness of noise is many times underestimated or even misunderstooded.   Pity that these AWG software made noises are  mostly far from ideal random gaussian noise.
How you characterize filter. Whos is using sweep generator. Other and very good way is use just noise. Input wide BW random (white or nearly white or bit pink) noise to filter and look what come out, thats it and so simple.
Even if you transmit RS232 over cable between equipments, how uoy know it work in real word if you have tested only with ideal "school book" signal. No, it need test in natural worst caase environment or you need simulate this situation, noise is one good tool for it. So, modulate your pulse queye with noise and know more how it work.

If we use signal generator only connected to oscilloscope and we want look nice images on the scope screen...  we want clean signal many times. But for real testings we need also dirty or very dirty signals and so that all parameters are in "worst case" position, including levels, freq tolerances, noises and mismatching ringing etc. These can simulate with good tools. Going to natural worst case real world testing is next phase but this IS expensive. One test may cost more that couple of equipments. ;)

(I have not tested SDG2000X noise modulation but SDG5000 I have tested and if talk seriously. It is NOT true gaussian random noise. It is some kind of "looks like random but not at all random" fake noise.

One example here. I can ask. is this random gaussian noise. No, it is far away. (yes there is trick and note also scale but... if look this with spectrum... it is fun)

If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 
The following users thanked this post: tautech

Offline tautech

  • Super Contributor
  • ***
  • Posts: 16025
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #396 on: May 31, 2016, 07:10:20 am »
It actually doesn't make any sense, if I were Siglent I'd remove that option in the modulation menus. They are looking for trouble by doing it this way.
This is what's being considered. ^^

Quote
However, adding noise to other waveforms (ie sine+noise, pulse+noise, etc...)  is very useful, as it lets you mimic real life conditions at the input of devices like logic gates, drivers, demodulators..... That feature should be added in my humble opinion. ^-^
Thanks for your feedback.  :-+
Avid Rabid Hobbyist
 

Offline Steps

  • Contributor
  • Posts: 23
  • Country: ru
Re: The Siglent SDG2042X Thread
« Reply #397 on: May 31, 2016, 01:47:50 pm »
Guide for Dummies

1. Plug in ethernet cable and turn on unit
2. Press Utility -> Page 1/2 -> Interface -> LAN Setup -> DHCP ON
3. On you computer telnet (Windows use Putty, OS X and Linux can just use telnet on the CLI) to the IP Address that's displayed. (If you need to manually specify the port it's 23)
4. mount -o remount,rw ubi2_0 /usr/bin/siglent/firmdata0
5. cp /usr/bin/siglent/firmdata0/NSP_system_info.xml /usr/bin/siglent/firmdata0/NSP_system_info.xml.orig
6. vi /usr/bin/siglent/firmdata0/NSP_system_info.xml
7. Use the down arrow to navigate to the line which has '<license><bandwidth_update_license>xxxx</bandwidth_update_license></license>'
8. Press dd to delete the line. Press ESC. Type :wq then hit enter (write quit)
9. Turn unit off and on again

Done!
Thanks. I feel myself like real hacker.
Sergey
 

Offline bson

  • Supporter
  • ****
  • Posts: 1554
  • Country: us
Re: The Siglent SDG2042X Thread
« Reply #398 on: June 01, 2016, 05:00:52 am »
Found another bug.  Set the output to a 5dBm sine wave.  Use the dial to reduce it in 1dBm steps.  It won't go negative.
Set the output to -5dBm.  Turn the dial to go positive.  It won't let you turn it back down to -5dBm again.  Once negative you can go up and down, but it refuses to cross 0dBm going downwards.
 
The following users thanked this post: tautech

Offline pascal_sweden

  • Super Contributor
  • ***
  • Posts: 1404
  • Country: no
Re: The Siglent SDG2042X Thread
« Reply #399 on: June 11, 2016, 11:53:05 pm »
High-performance sampling system with 1.2 GSa/s sampling rate and 16-bit vertical resolution.

Innovative TrueArb technology, based on a point-by-point architecture, supports any 8 pts ~ 8 Mpts Arb waveform with a sampling rate in range of 1 uSa/s ~ 75 MSa/s.

Where does the limitation of 75 MSa/s come from, while it uses a sampling system with 1.2 GSa/s?

For which signals does it use 1.2 GSa/s? And for which does it use less?

Why you would want to use a variable sampling rate that goes down to 1 uSa/s?
Isn't it better to just always use the highest available sampling rate for best results?

Or is it because of memory limitations, when you want to store longer periods?
But then again, isn't 1 uSa/s a bit on the low side? :)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf