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

[idpromnut]

   Right, so then you do have a later revision board than I (PCB 1002).

[trevwhite]

Hope this isnt noise but can someone confirm for me if the latest versions being sold are still hackable to upgrade the features?

Thanks

Trev

[idpromnut]

Wonder if that most recent firmware supports the HW I have, and if that firmware has support for the HF counter in it. Can you check (once modded) that you have the option to switch between the LF and HF counter in Utility - > Counter?

[marabut]

Hope this isnt noise but can someone confirm for me if the latest versions being sold are still hackable to upgrade the features?
Thanks
Trev

Yes, it is noise : I've mentioned what the difference is (in reply #254) and yes, it's still hackable 

Wonder if that most recent firmware supports the HW I have [...]

I have no idea -  take into account that there's no information concerning firmware upgrade availability on Hantek's website, so we don't know if there will be different firmwares for different board revisions. Personally, I doubt it (as they never did such thing to their DSO's).

[...] and if that firmware has support for the HF counter in it.

The board (and the firmware) should be the same for B and C models, regardless of revision. That's why they limit features through software settings - to rid off production costs for second, almost identical product.
And, after all, that's why the device is hackable.

Can you check (once modded) that you have the option to switch between the LF and HF counter

If there weren't "HF counter" option , I wouldn't be able to check if it works without U25 or not, right?

[tinhead]

marabut,

it would be great to see high res picturs of your board

[Cyber7]

@marabut: There's still plenty of old LMX5080 stock in Asia. You can find it on Alibaba, but my company doesn't trade on their supply xchange. I have found it at some distros I deal with, but their minimum is often >$300USD. I got a quote from one vendor at $50 each! I have some feelers out to some of my low vol suppliers for 5-10 pcs. If I get lucky, I'd be happy to supply the part.

I sure hope my unit doesn't come with heat sinks bonded to 3 soic-8 parts with silver conductive-epoxy. Looks like a post-engr eval unit to me, otherwise they would have revised the new PCB with proper stress mounts for the heat sinks. Shipping stress alone could rip the parts and pads right off the PCB...ouch! Frankly they should have redesigned the Faraday cans for airflow, lowered the profile above those 3 ICs and placed thermal coupling material in between. Even old C-64's and Apple II's used this design.

There doens't appear to be any thermal compensation on the outputs. Perhaps the output ICs get too hot with the cans in place. It would be interesting to see thermal images of the Faraday cans after 30mins of operation. Someone posted some thermal images of the regulators on the PCB running hot (albeit without forced air from the PSU's  fan).

[idpromnut]

Can you check (once modded) that you have the option to switch between the LF and HF counter

If there weren't "HF counter" option , I wouldn't be able to check if it works without U25 or not, right?

I suppose that kinda makes sense

[marabut]

@Cyber7: I know it is possible to order large quantity of LMX's - but it is not viable option (for me at least).  And I think that engineering eval board would have all parts populated, so probably it's rather an example of budget driven engineering.

@tinhead: how could I refuse the man who started it all...

Links to full sized pictures (warning: about 40 MB each!)
Top:
http://i1309.photobucket.com/albums/s629/lionhorse5/top_zps764ba714.png~original
Bottom:
http://i1309.photobucket.com/albums/s629/lionhorse5/bottom_zps7f9b67e3.png~original

Previews attached below:

[ulrik]

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

I got mine 2 days ago - and it's the very same version as marabut's. And I can confirm "upgradability" to HDG2102C regarding bandwidth    BTW: Thx to all upgrade-contributors! 

HowTo bandwidth upgrade:

• open case (very easy using torx)
• connect your TTL-UART next to USB connector - from USB: unsoldered, UART-TX, UART-RX; GND; 3.3V(not needed)
• use your favorite terminal prog - use setting 115200Baud,8N1
• switch on your Hantek device
• edit system.inf in /config/root (use vi-editor and change "HDG2002B" to "HDG2102C"
• then use "reboot" to restart system; done.

[tinhead]

@tinhead: how could I refuse the man

hehe, at average, one of my wives is always "refusing" 

-tcxo (or maybe OCXO?) is covered with cap / thermal insulation box

i think there is tcxo and not ocxo soldered (one can see on bottom, these 4 pins from ocxo pinout not soldered).
But if so, that box is of couse still nice to have.

Links to full sized pictures

cool, they nice, one can now easy compare to other models, especially these hand soldered parts, e.g. why D5 removed (near R238 / XO box), or why they soldered !2 when the P6/P7 (near usb on top) are anyway not used?

[Cyber7]

@marabut: Good News! I have qty 5 LMX5080M coming from China in 3 weeks. $12USD shipped

You can order from WWW.UTSOURCE.NET ...hopefully, this broker sells real, non-remarked parts....


Thanks for the pics. Interesting that VRegs U1 & U15 are missing, but they left the Vo filter caps behind weird.

Tinhead is right..through-holes for the OCXO source are unpopulated. Must be a TXCO under the shield cap. OCXO's may have greater stability (ppb), but don't you have to then wait for the entire unit to reach temperature equilibrium before trusting the OCXO output...

Also, an OCXO source would require a temperature stable reference voltage. Likely the missing U16 device immediately beneath it + assorted Res/Cap/diode. Probably something like a MAX6195/MAX6198. The foot print is right, but I don't think it's the right pinout as per your image & DS. Using an external 10Mhz rubidium timebase may be a better route, anyway.

[marabut]

@Cyber7 :
Great news, indeed! I've just sent an inquiry to them.

By the way, output stage of HDG doesn't like non-resistive loads. With some capacitance connected to output (1m of coaxial cable is enough) square wave looks OK until you add some DC offset - then things get worse. Much worse
I haven't seen such ringing for years.

[marabut]

@Cyber7 :
Great news, indeed! I've just sent an inquiry to them.

... and ordered the same quantity (but 14$ shipped)

[...] output stage of HDG doesn't like non-resistive loads.

It seems that problem appears when higher output voltages are set (starting from 4..5Vpp). When output signal amplitude is  increasing , there is a point where amplification is changed (you can hear relay switching). The very moment  depends on frequency (~4Vpp near 20Mhz, ~5Vpp @ 10MHz). Before switching occurs square wave is clean (more or less), after that... well, take a look at second screenshot.
Less dramatic than I've thought initially but still it's good to remember this limitation. 

[Stevie1966]

I have resolved this problem 3 days ago.
You can try the following.
Take away the capacitor C231 and C208. Then change L25 and L17 to 0 ohm.
Wish you luck. 

[alex.forencich]

I just got a new HDG2002B, and it seems the system.inf file has been moved from /etc/ to /config/root/system.inf. 

Software: 1.00.2 (140819.0)
Kernel: Linux 3.2.5
FPGA: 12
Keyboard: 3
PCB: 1004

Also, the DAC is not obscured: it is an Analog Devices TxDAC AD9747BCPZ.

I also tested it with python-usbtmc, and it seems to work.  I have not yet tested it extensively, though.  I will test the network connectivity with python-vxi11 as soon as I can get the parts to install the ethernet jack. 

[Stevie1966]

 :bullshit:AD9747BCPZ,very good.

 high dynamic range, dual digital-to-analog converters (DACs) with 16-bit resolutions and sample rates of up to 250 MSPS.

http://www.analog.com/en/digital-to-analog-converters/high-speed-da-converters/ad9747/products/product.html

[flatlander]

Got my HDG2002B today. System info below, FPGA version seems to have jumped a few versions.

Software: 1.00.2 (140926.0)
Kernel: Linux 3.2.35
FPGA: 20
Keyboard: 3
PCB: 1004

[idpromnut]

Got my HDG2002B today. System info below, FPGA version seems to have jumped a few versions.

Software: 1.00.2 (140926.0)
Kernel: Linux 3.2.35
FPGA: 20
Keyboard: 3
PCB: 1004

Nuts, now I really wish I could get a firmware upgrade for my unit 

[Cyber7]

Mine arrived last Saturday. I don't have the 5080 for the hi-req counter yet, but I've added the Ethernet components successfully. Ping works! Of course, there is only bare transport support at the moment. Now, maybe I can work out a proxy ala VXI for SCPI commands to the USB IF. Hmmm.... I'd love to use it with Agilent IO via tcp. Haven't tried the direct route yet via USB using Agilent drivers. Anyone have any luck this route?

Software: 1.00.2
Kernel: Linux 3.2.35
FPGA: 14
Keyboard: 3
PCB: 1004

 I am keeping a blog at http://hdg2002b.simplesite.com . So far I have consolidated all documentation I have: datasheets, hi-res scans & pics, scripts and upgrade procedures sourced and credited to the fine folks that have contributed to this forum. 

The firmware partitions for 1.00.2 are not the same as 1.00.1. They added a recovery partition. I updated the nanddump script for the new geometry, and succeeded with dum pin mtd0 to mtd8. I have successfully JTAGed the s3c2416 and NAND with a cheapo 100ask.net USB adapter via openocd. The scripts for USB nandbackup and the openocd config for the AWG target is attached below, and posted on the above site. Rename the .txt file as .cfg.

I also removed the heatsinks on the outputs and identified the part as a TDS3091. I've worked out a schematic of the output on Multisim...though, it's not yet functional.  I've made some assumptions on caps/inductors. I applied Stevie1966's fix for the outputs. I can confirm that it does remove the ringing from the under-dampened input to the TDS3001 amp stage at >=4v output.  THANKS! I've yet to verify that it hasn't altered the amplitude, but it looks good, albeit somewhat over dampened.

[Cyber7]

@FLatlander: Could you post your FPGA bin file? I'd be interested to see if they have done anything about the jitter that gets really bad above 5mhz for the standard waveforms.

Insert a fat32 USB stick and copy the bin file to the stick via this shell command over the serial port:

cp /dso/lib/firmware/* /mnt/udisk

zip the htg*.* files and post it here.

I've attached mine below. Note that the binary header for the file contains some useful info, including the device family "6slx16csg324", which is a Spartan-6 LX FPGA (XC6SLX16-CSG324) and the compilation date/time: 2014/06/15 10:01:28

"AFG3050.ncd;HW_TIMEOUT=FALSE;UserID=0xFFFFFFFF.b..6slx16csg324.c..2014/06/15.d..10:01:28"

[flatlander]

@FLatlander: Could you post your FPGA bin file? I'd be interested to see if they have done anything about the jitter that gets really bad above 5mhz for the standard waveforms.

OK, copied the FPGA firmware file (see attachment). The date in the file header is 2014/09/20, so a few months more recent than Cyber7's. I also unlocked the extra bandwidth, which still seems to work but I haven't had time to do extensive testing.

@Cyber7: Great initiative with the HDG2002B website, looks great. Good to have everything easily accessible in one place.

[FrankenPC]

That website is great!  What does the audio section theoretically do?

[Cyber7]

@flatlander: Thanks! I'll try it out tonight.

@FrankenPC: My guess: Hantek copied all the peripherals from the devkit! Really, they probably thought to provide some bells and whistles, but a piezo buzzer is much cheaper than a codec, amp, spkr combo.

Maybe some 'enterprising' engineer out there will work up an LCARS skin and audio mod!

I'm presently digging around LXI & VXI-11 specs/codes for linux to get familiar with the protocol. I may be able to get away with a sockets server daemon that pipes messages to the Android USB driver, as the SCPI language support looks fairly extensive based on the English user manual. I have a 2416SDK as well as the GPL sources for the DSO5000 (same kernel rev, and similar U-boot), so I'll see about adding FTP, SSH, etc, to simplify mod development.

[Cyber7]

Ok, while I was testing the DM9000 ethernet mod, I ran the AWG's PCB outside the case without the front panel, with a separate cooling fan across the PS and main PCB. So I used TTSOURCE to control the unit. No problems. Now that I have put everything back together: The front panel controls the device, LCD works, self diags all OK, USB ok, ethernet pings + telnetd mod works great. TTSOURCE reads the unit and controls the AWG, BUT the LCD & button LEDs no longer update when I make a change in TTSOURCE! 

I've reset the AWG firmware to factory defaults using the U-boot recovery partition. Lost my software tweaks and BW limit mod, but still no dice; the front panel will not reflect changes made via USB/TTSOURCE. I've also had a few BSODs on the win64 host PC while changing controls via TTSOURCE.

Tried, with No effect: Firmware reset to factory settings, ethernet cable jacked in/out, different USB cable, different PC & driver.

Hardware mods: Ethernet (added all missing parts), Hi-freq counter (added 5080  IC and RF jack), output fix (removed caps, inductors).

 Anyone with the DM9000 mod have a similar issue?

EDIT: Guess I was imagining the display/front panel updated when TTSOURCE was modified. Just reread the thread, and idpromnut wrote last July regarding TTSOURCE control:

You won't see anything on the HDG itself as the leds and the display are not updated but you will hear the relays when switching a channel on or off.

 

Can anyone confirm this behavior with a unit running FW v1.00.2 ? Thx.

[Cyber7]

In my apparent Quixotic attempt to 'fix' the ttsource/front panel issue I did find this interesting bit:

SCPI Commands are found within library: libhtg1000.so
used by:
test_cmd
afg_bin

[root@Hantek /dso/app]#./test_cmd
./test_cmd: error while loading shared libraries: libanolis-1.1.0.so: cannot open shared object file: No such file or directory

Argh. Ok let try simlinking to the 0.0.0 lib available:

ln -s /dso/lib/libanolis.so.0.0.0 /dso/lib/libanolis-1.1.0.so

-------------------------------------------------------------------------------
./test_cmd
lib htg version:1.00.1.0
<7>dma0: s3c2410_dma_request:734: ls=0, cur=  (null),   (null)   (null)
0
Usage: htg -cmd `scpi cmds`
0

Aha! it runs. Funny thing is that I hear relays clicking even when no command is specified.
-------------------------------------------------------------------------------
./test_cmd -cmd *IDN?
lib htg version:1.00.1.0
0
input is *IDN?
findCommand: cmd founded, *IDN?
hantek,HDG2002B,DDSxxxxxxx,result is 1

serial xxx'ed out

-------------------------------------------------------------------------------
./test_cmd -cmd IDN?
lib htg version:1.00.1.0
<7>dma1: 00000002 to DMASKTRIG
0
input is IDN?
SCPI_Parse: undefined cmd
**ERROR: -113, "Undefined header"
result is 0
0

Purposedly skipped the * so its invalid.

Interesting thing is that sometimes when turning a Output On/OFF I would get:
SQL error: database is locked

Seems the control's status is maintained via SQLlite.