Hantek HDG2002B AWG: 5Mhz or 100MHz? Let's see!

[marabut]

>>could the HDG2002 generate sine wave about 25 MHz?
Out of the box -> no.
After modding -> yes.

[ulrik]

Arbitray waveforms file format is .arb, and it can be opened by .txt.

So is it a similar data format like i.e. comma separated values (csv) or tab delimited values? Please, could you give an example of such a file (at least a header)? .txt format does not ensure its a proprietary data format. Thx!

[marabut]

fremen67 posted some waveform files here (see page 14 of this thread), so you can download one and check  file format on your own. 

[rosasharn]

Arbitray waveforms file format is .arb, and it can be opened by .txt.

So is it a similar data format like i.e. comma separated values (csv) or tab delimited values? Please, could you give an example of such a file (at least a header)? .txt format does not ensure its a proprietary data format. Thx!

Sorry, there's a mistake in my last post. The data format for HDG2000B is .hwf.
The content of this .hwf file is csv file format, and it's easy to compile data by software like excel.

Example

version,17
type,2
samples,128
sps,250000000.000000
zoom,20
0.004570
0.042962
0.074040
0.088665
0.108775
0.138026
...

First line 'version,17' stands for the file content format version is 17, no need to modify;
The second line 'type,2' stands for the record data type is floating point data, its range: -1 to +1 (corresponding voltage range: -0.625V to + 0.625V). When type is 1, the record data type is integer, range: -32768 to +32767 (corresponding voltage range: -0.625V to + 0.625V);
The third line 'samples,128' stands for waveform data points is 128 points;
The fourth line 'sps,250000000.000000' means sampling rate is 250000000hz, i.e. the interval between two points is 4ns;
The fifth line of 'zoom,20', means the waveform magnification is 20 times, effective numbers are 1,4,20, stands for the corresponding magnification for the voltage of waveform;
The rest lines, each line represents a waveform data.

[marabut]

Some fresh news:
-currently shipped HDG2002B have firmware version 1.00.2(140819.0) , PCB: 1004, FPGA: 12, Keyboard: 3, Kernel: Linux 3.2.35
-U25 place on PCB is not populated, so HF counter doesn't work and upgrade becomes more complicated
-system.inf is now located inside /config/root
-tcxo (or maybe OCXO?) is covered with cap / thermal insulation box

That's all for now. Now I need to find proper replacement for U25 
What exactly is U25 marked? Is it a common HF divider?

[Cyber7]

Tinhead suggests, as per the 1st post, that U25 is an LMX5080 or a MB510 ( 2.7 GHz Low Power Dual Modulus Prescalers).  U25 is definitely made by National Semiconductor, but I can't read the part # in the lo-res image posted by fremen67. There is also a non-populated SOIC8 part on the bottom of the PCB near the non-populated RF jack. Perhaps, the above reference is to this IC instead? Anyone have a good pics of the top/bot of the PCB in this area? 

Problem with the LMX5080 is that it is no longer in production and difficult to source. The Fujitsu MB510 is pin compatible, however, it only implements 4 divide ratios, while the 5080 utilizes pin 7 to double the # of ratios. There  ratios of  1/128 and 1/256 are the only ratios they share in common. 1/512 is not available on the MB510. 2.7Ghz / 256 is ~10Mhz, so 1/512 may not be required as a prescaler.

I should have my HDG2002B in 2 weeks, so I'm hoping I won't have the same issue.  I'll put the PCB on an old optical scanner and get some hires images.

I've attached a lo-res image of the u25 area populated, and a LMX5080 SOIC8 in someone's Faraday cage adapter.

[leppie]

I am still tempted to get one of these. Have some spare cash that I saved from getting a DS1054Z

Is it worth it?

[idpromnut]

I am still tempted to get one of these. Have some spare cash that I saved from getting a DS1054Z

Is it worth it?

For the ~300$ you are going to spend, if you want new then yes, it's a pretty good deal for what you get. Like always, you can probably find a better deal on EBay if you wait a bit.

[leppie]

I dont really need it per se. Maybe I should look for a better bench supply as I arguably have pretty much the worst bench power supply in the world (ATTEN APS3005S). You ask why? Hundreds of volts overshoot when powering on. Got into a habit of connecting to positive terminal only after power on.

[idpromnut]

Hmm, I don't know about a new bench supply, but I know you definitely have a Repair Project on your hands!

[leppie]

Hmm, I don't know about a new bench supply, but I know you definitely have a Repair Project on your hands!

I will wait for the new scope, so I can get some screen caps. Taking photo's on the TDS360 is troublesome at best

[marabut]

Tinhead suggests, as per the 1st post, that U25 is an LMX5080 or a MB510 [...]

Is it possible they're using obsolete parts in quite contemporary design?
I've searched (not thoroughly) NS(now TI) website but I couldn't find any pin-to-pin replacement part being in production now. MB510 is easier to find, there is a few offers on eBay now.
According to datasheets, it seems that with pins 3, 6, 7 tied to Vcc both parts should have 1/128 divide ratio (however for MB510 pin 7 is described as NC and should be left unconnected). LMX is more sensitive though (-15dB vs. -10dB) and can work with 2.7V Vcc (MB510 - 5V only).

Maybe somebody (who has unit with working HF counter) can check U25 marking and post the info here? 

[idpromnut]

Tinhead suggests, as per the 1st post, that U25 is an LMX5080 or a MB510 [...]

Is it possible they're using obsolete parts in quite contemporary design?
I've searched (not thoroughly) NS(now TI) website but I couldn't find any pin-to-pin replacement part being in production now. MB510 is easier to find, there is a few offers on eBay now.
According to datasheets, it seems that with pins 3, 6, 7 tied to Vcc both parts should have 1/128 divide ratio (however for MB510 pin 7 is described as NC and should be left unconnected). LMX is more sensitive though (-15dB vs. -10dB) and can work with 2.7V Vcc (MB510 - 5V only).

Maybe somebody (who has unit with working HF counter) can check U25 marking and post the info here?

What makes you think you do not have a working HF counter?

[Cyber7]

Is it possible they're using obsolete parts in quite contemporary design?

Possibly, considering that they have nixed the feature. You can still find stock of the 5080, but only in lot qty from overstock/grey suppliers. HT probably had spare stock left over from the DDS3005 (2.7GHz Frequency Counter, PC USB Function/Arbitrary Waveform Generator). Looks like the same LF: DC to 25Mhz / HF: 25Mhz to 2.7Ghz design params.

[Cyber7]

idpromnut:

What makes you think you do not have a working HF counter?

In Reply #237, you posted pix of your HF counter mod, including a populated U25 adjacent to the jack you replaced.
marabut's unit is not populated for U25 as per his pix. Looks like fremen67 had the same issue based on his initial postings, and tinhead's reply: "the Frequency Counter, you right, that will be something like LMX5080 or MB510". If U25 is the HF prescaler, then attaching a lead to the PCB directly with a test osc should fail to register anything on a 'upgraded' unit, right?

Do you have a better pic of the U25 area, or happen to make note of the part #?

I won't have my unit delivered for another 2 weeks....so I'm just fishing based on the pix and posts. Also trying to get a leg up, in case I need to populate the part too.

[idpromnut]

Do you have a better pic of the U25 area, or happen to make note of the part #?

Nuts, you're right!  I pulled apart my unit just now to check and indeed, I do have a SO-8 in U25, and yes, it is a LMX5080.

[Cyber7]

  Many thanks for the sanity check!

It's gonna be a PITA to get an LMX5080 if mine is missing...

Sadly the MB510 if a 5v part and Vcc is 3.3v. Worst case, the Ti campus is only 15 miles away; I could go up and ask to see someone in their RF comp section and beg/horse trade for old stock samples...LOL

[marabut]

idpromnut, I've checked HF counter exactly the same way fremen67 did:

I tried to connect one  channel  set to the max frequency (100Mhz Sine) to the HF counter directly on the PCB and guess what... it worked! It seems that everything, except the jack, is already on board for the HF counter to work 

Thank you very much for checking U25  .
Now I know what to lok for, at least.

Cyber7, I hope you will have more luck than me. I can't find LMX5080 anywhere near my location, even eBay search returned no result (not a single one, literally!). On the other forum, somebody suggested to look for old mobile phone / Tsat tuner boards where LMX's were used. I will ask on HAM radio forum as well. Maybe they have some...

By the way, I have tried to measure 10MHz reference output (using LF counter channel). I've expected exactly 10 000 000 as the same clock is used to control gate circuit, but result was somewhere near and changed with almost every read... Counter design flaw? Gate time generated by software?
Strange behavior, anyway.

[idpromnut]

By the way, I have tried to measure 10MHz reference output (using LF counter channel). I've expected exactly 10 000 000 as the same clock is used to control gate circuit, but result was somewhere near and changed with almost every read... Counter design flaw? Gate time generated by software?
Strange behavior, anyway.

It definitely feels like a software-based counter. After reading this, I did a couple more tests with the HF counter, and noticed that the HF counter seems to have sub-par termination. Testing with a 10MHz 5Vpp sine wave, we get this when the FG is connected to the LF counter (HDG is reading 9.882MHz):


And this when connected to the HF counter (HDG is reading 9.879MHz):


With a 15MHz 5Vpp sine wave, we get this on the LF counter (HDG is reading 14.762MHz):


And this when connected to the HF counter (HDG is reading 14.761MHz):

[marabut]

I think you shouldn't drive HF counter input with anything above ~1Vpp (unless you want to fry input circuit components  )
From PCB traces (take a look at the picture I've posted)  you can see that signal is routed from input connector through small capacitor to two antiparallel connected diodes D9 (signal limiter) and then through another capacitor to LMX input.
LMX5080 can work with maximum signal +4dBm (according to datasheet) which translates to 1Vpp. Even "Absolute maximum ratings" say that voltage on *any* pin cannot exceed Vcc+0.3V. I don't know actual Vcc but it's more safe to assume that Vcc is 3V rather than 5V.
So, feeding it with 5V signal doesn't serve it well

It is possible that sinewave distortion you see is caused by limiter circuit. With smaller (<1Vpp) signal waveform will be better, I guess.
 

[idpromnut]

Fair enough.  I did another quick test and yes, @ 1Vpp the HF counter doesn't seem to distort/attenuate the signal. Wackiness I tell you 

[Cyber7]

Marabut, seems you are also missing the shielding around the outputs, and have green heatsinks on several IC's ? 

Is this factory rigging, or are you having output stability issues?

[marabut]

There was no shielding around the output stage and heatsinks are original.
I've removed only external shielding to take pictures (and to connect serial cable of course  ).

[idpromnut]

There was no shielding around the output stage and heatsinks are original.
I've removed only external shielding to take pictures (and to connect serial cable of course  ).

Can you post the board revision? I wonder if the unit I have is a older or more recent one than yours. Mine is missing those heat sinks and has the metal cans on the output stages as well.

[marabut]

Can you post the board revision?

Of course  :

Some fresh news:
-currently shipped HDG2002B have firmware version 1.00.2(140819.0) , PCB: 1004, FPGA: 12, Keyboard: 3, Kernel: Linux 3.2.35
[...]