Hantek 4032L Logic Analyzer

[Fraser]

I have been looking for a new Logic Analyzer and, after some searching, found that the Hantek HT4032L meets my limited needs for monitoring data and address buses, at a decent sample rate. 32 Channels, 2Gbit internal DDR2 memory, up to 400Ms/s, 150MHz b/w.

It is a USB device so uses a PC, but it contains 2Gbit of DDR2 Ram that is used across the 32 Channels as required 

I am aware that Hantek software can be flakey but hopefully it will be good enough for my requirements. The SDK is listed on the Hantek web site, so DIY software is a possibility.
I shall report my thoughts on the unit when it arrives this week. It is available in the UK from Rigoloscilloscope.co.uk for GBP139 delivered.

http://www.rigoloscilloscope.co.uk/wholesale/PC-USB-Logic-Analyzer-Hantek4032L-320.html

Manual attached

[Lightages]

I hope to see a review soon! Please?

[Fraser]

The 4032 arrived today. It was taken apart within minutes 

Here are the pictures of what lives inside. The PCB is very well anotated, almost as though Hantek are encouraging experimentation with it. Even the JTAG connection is clearly maked, complete with voltage pin !

On the down side, when I received the unit something just didn't look right. In the all the pictures that I had seen, the annotations on the I/O socket were upright when the unit is stood vertically on its feet....not on my unit, they were all upside down. To overcome this I took the unit apart and changed the orientation of the case so that the anotations are upright. The Hantek logo is now correctly orientated but on the opposite side to the pictures on the internet ! The power and USB ports are now nearesr the top of the unit rather than near  the desktop. A small FAIL, but still a FAIL in my book. The label on the bottom also reads "2M Memory Deepth". A minor spelling error but a silly one for a manufacturer to make as all labels should be proof read before release.

The port marked USB-X1 is, I believe, a port used to connect the unit to an oscilloscope and other Hantek test equipment to form a test kit 'stack'

The PCB is marked "2012-12-7 LA4032 V3.0" so it has been through at least two other versions and is a relatively recent PCB release. The PCB and the component placement is very good, so no complaints there. Just how could they have messed up the I/O port orientation so badly !

The PCB design seems a step up from the Hantek 5034 which, IMHO, looks to be a far cheaper unit in terms of quality of layout. The 4032 has nice track length compensation (the wiggly pcb tracks) and a well organised layout. Compare my pictures with those of the 5034 in this thread:

https://www.eevblog.com/forum/chat/anyone-tried-the-hantek-la5034-logic-analyzer/msg18290/#msg18290

I wouldn't have thought that they were both from the same OEM manufacturer but I may be wrong about that.

The logic analyzer package includes the 4032L, A decent looking multi national SMPSU with US, Aus and European adapters (no UK adapter), nice quality short length probe leads, 18 IC pin grabbers of reasonable quality (I will be using the superior HP and TEK ones though), USB cable and a software CD.

All in all I am pleased with the design and build of the unit. I have an HP logic analyzer self paced training board that I will connect this unit in order to do some tests on its performance.

Chip data:

Spartan-6 XC 6SLX16 :

http://www.xilinx.com/products/silicon-devices/fpga/spartan-6/lx.html

Cypress CY7C68013A-100AXC :

http://www.cypress.com/?mpn=CY7C68013A-100AXC

Micron D9LHT:

1Gbx16 400MHz SDRAM DDR2

http://www.micron.com/parts/dram/ddr2-sdram/mt47h64m16hr-25e

[nctnico]

IMHO its pretty sad the designer used so many tantalum capacitors. These should be avoided because they are unreliable and not so good compared to ceramic multilayer capacitors.

[Fraser]

Noted but is the Tantalum not maligned because of its failure in old equipment ? It has always been used in MilSpec equipment made by Racal and they served their country reliably until they reached an age well beyond that intended for continued use. I have repaired many Racal equipment's with failed tantalum capacitors on the power rails, but that equipment dated back to the 1980's. I am suspicious of Electrolytic capacitors even in relatively new equipment but I was not of the same opinion about Tantalums....am I wrong to trust them more than the crap cheap electrolytics ? Tants have always been the expensive and preferred alternative to electrolytics.

[nctnico]

Because of the low ESR they are better for switchers and power supply decoupling. The whole problem with a tantalum is that it is basically a small bomb (oxidizer and fuel). A large voltage spike may set it off and if the PSU can supply enough power it will burn right through a board like thermite otherwise it will just become a short.

[Fraser]

Ah I see why you are not a fan of them. I have been lucky to date, only shorted tantalums that took out the power rail rather than burning anything. Nest time I restore a piece of old kit I think the Tants (good or bad) will be changed out.

Thanks for enlightening me on these little monsters.

[notsob]

I think it comes down to using the 'right' component in the right place and beware of it's possible short comings when designing. I've repaired boards where many of the resistors were ' fuse' varieties, ie failed open to prevent possible damage, and at the same time boards with gigantic holes burned thru them when a tant went up in flames.

[Lightages]

Thanks Aurora. Now we need to see it in action! 

[codeboy2k]

That Cypress chip is a pretty expensive USB controller; it's around $11.00 for qty 1000+... I think if I was doing it today, I might consider an ATMEL device with USB. They can cost  as little as $5.50 or so in qty 1000. It would really depend on the availability of DMA on the chip and USB software libraries for the device.  But it might be possible at half the cost, unless they would have to license something that the Cypress chip includes already.  If so, well then there's the cost difference right there. .

They are probably using Cypress's well known and well supported EZ USB

[Fraser]

I was pleased to see known brands of chips in the unit and I am also pleased to hear that they have used a relatively expensive Cypress chip. It may not be the most cost effective solution but I hope this means that the unit is not a cheapo, built down to a budget product using the very cheapest parts available. I am also pleased to see that no silly chip anonymisation has been done !

Thanks for the comments on the design, most welcome. I only hope the software is worthy of the hardware !

I don't have time to 'play' at the moment, so pictures of the software working will need to wait a little.
 

[marmad]

I was pleased to see known brands of chips in the unit and I am also pleased to hear that they have used a relatively expensive Cypress chip. It may not be the most cost effective solution but I hope this means that the unit is not a cheapo, built down to a budget product using the very cheapest parts available. I am also pleased to see that no silly chip anonymisation has been done !

It's a pity they didn't provide a Trigger Out - very little extra cost to do it - and it really should be standard with LAs.

Thanks for the comments on the design, most welcome. I only hope the software is worthy of the hardware !

Not to jump the gun on your review, but the software looks exactly the same as the software for their old LA (LA5034), which hadn't gotten very good reviews by a previous owner here. Oh well, Hantek software (and you can say the same for pretty much all Chinese test equipment manufacturers) is generally crap - so you just have to resort to writing your own.

[tinhead]

I was pleased to see known brands of chips in the unit and I am also pleased to hear that they have used a relatively expensive Cypress chip. It may not be the most cost effective solution but I hope this means that the unit is not a cheapo, built down to a budget product using the very cheapest parts available. I am also pleased to see that no silly chip anonymisation has been done !

It's a pity they didn't provide a Trigger Out - very little extra cost to do it - and it really should be standard with LAs.

i bet this USBXI port provide trigger out as it can be used to sync multiple devices (as far i understood).
The problem is, i don't see any pinout nor documentation published 

[marmad]

i bet this USBXI port provide trigger out as it can be used to sync multiple devices (as far i understood).
The problem is, i don't see any pinout nor documentation published 

Maybe, but even if true, it's certainly not handy on a card-edge connector.

[JimmyMz]

I guess there isn't much to say about this LA, or am I reading the 2+ weeks of time w/o a review poorly?

[Fraser]

Reviewing equipment is not my day job and so other priorities come first.

Patience will be required by those wanting a review as these are just 'toys' to play with compared to my day job responsibilities. There are never enough hours in a day 

[JimmyMz]

Personally, I don't need a lengthy review. A thumbs up, or thumbs down, along with two sentences; one sentence for pros and one sentence for cons, and I'd be happy. 

Please don't misinterpret my comments, I fully understand you might be busy, and I'm completely happy, if you write something once you feel like it.

Thank You,
JimmyMz

[Fraser]

OK, I will sort something out this week for you. It will not be a lengthy review but will detail my findings when used on a digital test piece that I have. I also have the Hantek 1025G Function/Data pattern generator that will present the LA with some different data patterns. I really want to get stuck into testing these but alas life has been on the hectic side of busy since purchase.

For info I also have a Sysclk-DX en-route to me. That has a lesser sampling frequency capability, but does offer data decoding and two low frequency oscilloscope channels in the one box. A true MSO it isn't but it may be an interesting 16 channel unit all the same.

Sysclk-DX Specifications:


Two input channels with a maximum analogue sampling rate of 24Msps
   Analogue Bandwidth 3 MHz
   Analogue Channels: 2
   Input voltage measured: -10V to 10V
   Input impedance: 1MOhm/30pF
   Logic I/O Channels :16 


   Protocols:
              Channels: 16

   Maximum sampling rate: from 1 to 24Msps with 4 trigger levels
   SPI, I2C, Async, USB, PS/2, SM Bus, Sync Serial, CAN, 1-Wire, Parallel busses as well as Custom Decoder capabilities 

Utilities:
   Bus Data Extractors
   Oscilloscope/Logic Analyzer
   Bus Decoders
   PacketPresenter
   Digital Signal Generator
   Data Logger
   Voltmeter
   Frequency Counter
   I2C Bus Controller
   Frequency Generator
   PWM Controller
   Pulse Counter
   Toolbuilder Source Code API
   Remote Controller


http://www.aliexpress.com/item/Free-Shipping-Sysclk-DX-Mixed-Signal-Oscilloscope-and-logic-analyzer-USB-2-probe-P6100-20LA-probes/583945355.html


I also own a Link Instruments 200Ms/s MSO-19 that is an 8 Channel MSO with one analogue input.

http://www.linkinstruments.com/mso19.htm

Its a very clever product but it has only 8 channels in the LA and cost a lot more than the other two LA's.

Watch this space.

[abyrvalg]

I'm another fresh happy [hard to say] owner of this device, so here are my 5 eurocents:

Hardware:

• + big buffer
• + build quality is much higher than expected
• + a good FGPA+DDR+USB plaform (more on this later)
• - clips quality is poor
• - plastic bumpers of the housing produce strong "chinese equipment" smell (thrown away immediately)

Software:

• + trigger options beyond simple single edge/single pattern
• + protocol decoders seems to work
• - stucks in "Capturing..." state often (Stop button does nothing, or there is some very big delay?)
• - no data save/load function, export only
• - waveform render is extremely slow! zoom/scroll operations takes seconds
• - using cursors is a bit complicated - right click, select from menu
• - no external clocking option (although mentioned in manual, ACLK/BCLK pins present on a connector, but Hantek support had confirmed that it's not possible)

I don't expect breakthrought sw updates from manufacturer, but hw part looks pretty powerful, so I'll try some DIY based on it (first - try adapting some 3rd party sw like open source sigrok).

I did a quick research already, here are some results:

• Both FX2 and FPGA boots to their final states from serial EEPROMs, not from PC (not so handy for experiments, but this can be fixed - disconnect both EEPROMs, switch Spartan to Slave Serial mode)
• FX2's fw is pretty dumb, data passes directly to/from FPGA with no sw intervention (slave FIFO mode)
• Spartan's JTAG pinout found (so I got access to spartan's pins via boundary scan)
• All FX2-Spartan interconnect found
• All Spartan-test port interconnect found
• Both DRAMs are connected to Spartan's dedicated MCBs
• The undocumented USBXI is FX2's USB, 2 Spartan's IOs (triggers?), power/gnd
• Test port has two reference voltages (defines treshold voltages) generated by obscured U11,U12,U13. U13 is connected to Spartan

If somebody wants to play with this device, I can post more details.

To codeboy2k: Cypress FX2 gives both simplicity and high throughput (20+ MBytes/s can be reached with a few lines of code, most of the hard work is done by hw). AFAIK comparable Atmel chips (ones with USB 2.0 HS ports) are ARM9s suitable for Linux, with much more complex programming.

[Rasz]

If somebody wants to play with this device, I can post more details.

Have you heard about http://sigrok.org ?
They are currently in a process of standardizing/modularizing logic sniffer firmware codebase so it can be reused on  all of the supported devices.
http://sourceforge.net/mailarchive/forum.php?thread_name=51923CBA.4000308%40ciellt.se&forum_name=sigrok-devel

[Fraser]

abyrvalg,

Thanks for your excellent insight into the 4032L. I will see if mine behaves the same with the lock-up issue and slow responses. I am running Win7(64) on an i5 so there should be enough grunt on the PC side for this little unit.

Out of interest did you notice that 'as supplied' the front panel legends are upside down. That is to say when the unit is stood up on its side with the word HANTEK facing you, all the I/O connector details are inverted ! An amazing cock-up ! I know what you mean about the rubber bumpers, but mine don't smell too bad at all.

I will be testing my unit against an HP MSO training board that provides all sorts of test signals to play with (GBP30 new on e*ay). I will also use the Hantek 1025G but I will need to be careful as I could end up with the scenario where poor output from the 1025 causes apparent issues with the 4032 

For anyone interested in a simple MSO training board and manual look here:

http://www.ebay.co.uk/itm/Self-Paced-Training-Kit-for-Hewlett-Packard-546000B-Oscilloscope-/290841237442?pt=UK_BOI_Electrical_Test_Measurement_Equipment_ET&hash=item43b77d17c2

and another sellers offering:

Silex-Major - he normally has a few new ones in stock

http://www.ebay.co.uk/itm/Hewlett-Packard-54654A-Self-Paced-Training-Kit-/370792531262?pt=UK_BOI_Electrical_Test_Measurement_Equipment_ET&hash=item5654f5053e&nma=true&si=HmEgAcW1wJEWok98evDCIhXKKFY%253D&orig_cvip=true&rt=nc&_trksid=p2047675.l2557

and here for the USA:

http://www.ebay.co.uk/itm/HP-AGILENT-54654A-TRAINING-KIT-/290750149799?pt=LH_DefaultDomain_0&hash=item43b20f34a7

Its pretty good for the money being asked and is a good quick start test piece for newbies.

[abyrvalg]

Rasz, that long link is very interesting, thanks! I've seen their top level pages, but never went deeper

Aurora,
Yes, my unit had the same mistake with panels, bumpers legs were facing opposite sides also I have a feeling that housing/clips/sw were designed by some different team than PCB.

Btw, I had some progress today, patched the board to boot from PC (boot FX2 with fpga loader fw, load fpga, boot FX2 with function fw), original fx2/spartan binaries boots just fine. The result is that LA behaves like a generic bootable fx2/spartan/ddr evaluation board with possibility to boot original LA function in 5 seconds.

[Rasz]

Rasz, that long link is very interesting, thanks! I've seen their top level pages, but never went deeper

Sigrok guys would probably appreciate it if you shared all the info about hantek with them. They might even help you with firmware if you want to extend existing feature set.

[Fraser]

With regard to the provided test clips..... I have found most cheap one to be a liability so have bought up plenty of the HP/Agilent and Tektronix micro grabbers. A far superior product, but they can cost a small fortune unless, like me, you find a reasonable seller. I have paid around GBP20 for a new pack of 20. They are commonly available in the USA but sadly the postage to the UK can be a real killer.

The additional cost of the quality grabbers is fully justified if you consider the potential damage that can occur if IC pins are shorted together by a cheap grabber claw. I elected to save a few dollars on the SysClk DX unit, by not having the grabbers included. No loss at all to me as they are pretty rubbish.

[nctnico]

Grabbers always have been a nuisance. The rate in which they detach per minute is e^(number of grabbers)
I usually solder thin (like hair) wires to the pins I want to observe. Those thin wires go to a header which I can plug into the probes of my logic analyser. An advantage of the thin wires is that they also work on 0.5mm (or less) pitched QFP and QFN packages.

[tinhead]

If somebody wants to play with this device, I can post more details.

as you did some reversing already, you could as well draw schematic of that LA 

[tinhead]

Test port has two reference voltages (defines treshold voltages) generated by obscured U11,U12,U13. U13 is connected to

and they are

U13 - DAC122S085CIMM
U11, U12 - probably LM358D

the circuit look very similar to what Hantek is using in their MSOs, see attached schematic

[hansan]

Dear all,

I have a question concerning this Hantek 4032L logic analyser.  Did someone try to get it working on Linux? 

It's hardware feature set, with the relative low price makes it a very nice tool. However my development environment at home  is linux only.
I have not a very high hope: Wine and USB (drivers etc) is in general not a good combination and I didn't see this as logic analyzer supported under sigrok either. But it would be nice if it could work under LInux.
Probably I have to settle for now for the Open workbench Logic Sniffer or are there more alternatives?

Greetings

[abyrvalg]

tinhead, good job! Glad to see you diggning into this device too 

Here are my results:

Protocol/hw info:
http://sigrok.org/wiki/Hantek_4032L
http://sigrok.org/wiki/Hantek_4032L/Info

Original fx2/fpga images and loader for "PC boot" modded hw (a generic Cypress FX2 driver is used): load_original_fw.zip attached
Loader sources: loader.zip attached
SPI eeprom reader with sources: read_spi_eep.zip attached
LA input pin schematics: input_pin_sch.png attached
Schematics netlists: netlists.zip attached
I had no time to draw a nice schematics (digital part's drawing is not of much interest to me anyway, as FPGA pin names say nothing), just made netlists of most important connections, must be enough to play with FX2/Spartan.

[cancian123]

Hi Guys. I bought a 4032L , and I have some problems with it. Turns out that after some minutes it just stop taking measurements (after around 10 minutes) and then I have to power it off and on again in order to make it work. I contacted the seller, and he said that it seems that there is a problem due to the big amount of memory. Hantek sent to me a firmware update, but the same thing is still happening. I would like to know if anyone here has the same problem.
Thank you.

[Scratch.HTF]

With reprogramming of the FPGA SPI flash (MX25L4005) via JTAG (/WP can only be controlled via JTAG - the flash cannot be programmed or even read via the SPI header even with the Reset jumper closed according to my experience), the following functions can be possible with modified host software:

* Data compression as in the SysClk LWLA1034 (and others)
* Qualifier capability (record data only on qualifying conditions)
* If-Then-Else sequencer with full conditional structure (up to a certain number of levels)
* Multiple state trigger words
* Range recognizer
* Selective data acquisition functions
* Programmable STOP conditions
* 64 bit (probably shorter) FIFO stream trigger (good for triggering on I2C and SPI)
* Option to combine logic ports to extend number of points (e.g. 32-16-8-4-2-1 bit for 64M-128M-256M-512M-1G-2G points)

From what I know, the Cypress FX2LP core has the ability to directly edit the data contents of the internal 16 KByte RAM and the internal 512 byte scratchpad RAM via a vendor specific command defined in the boot ROM for "soft" user code downloads.

JP2 (FPGA JTAG)

1   GND
2   TMS
3   TDO
4   TDI
5   TCK
6   +3.3V

JP6 (FPGA Reset - may need to leave unshunted to enable external reading/writing of FPGA SPI Flash via JTAG)

1   /RESET
2   GND

JP7 (FPGA Flash SPI)

1   +3.3V
2   SCK
3   SI
4   SO
5   /CS
6   GND

All pins have 4K7 pullups to +3.3V.

I will provide images of the Cypress FX2LP boot ROM (24LC64) and FPGA flash (which may be tied to a particular Xilinx DeviceDNA ID) on request.

[abyrvalg]

Scratch.HTF, WP# is on separate JP5 (close to drive high). All SPI signals are connected to FX2 also - this can be the source of your r/w problems. Reading the flash via FX2 works fine (see my sources above).

[Moskvoshveya]

I'm sorry, but have you any sample FPGA project, which is using memory, soldered on this board? I want try to make own firmware (of course, for add compression), but have no experience with Xilinx Core generators. Using already built DDR2 core will helps to save time by skipping MIG investigation.

FX2LP and other things are not a problem. Most problem is Xilinx Based tool because all my experience is Altera based.

[lfldp]

hello
have one question , does this hantek 4032L have trigger out functionallity ?
regards

[meandeev]

sorry for reviving this old thread, but I got a half-dead (doesn´t connect to the pc via USB). So I desoldered U3 (24L64, fx image), saved the old data and flashed the data from #28 with an external programmer. But after that the LED no longer lights up! And also no USB connection...

The LED´s works again after switching back to my saved data. I switched the data once more to be sure that I didn´t make an error. Same result.

If one compares both images, one can find missing byte here and there in the data from #28 - because of reading the data via USB?

So is there  an error in #28? The LED is driven by the fx (pin92=PE6 and pin91=PE5).

So here my U3 fx data is attached (desoldered and read with an external programmer)


PS: reading U8 (fpga data stream) with an external programmer give the same data as in #28

[B-Good]

I have the same content of the same situation.
with the new firmware "fx2_v0.bin" nothing works
with the original firmware I can't use PulseView.
it gives me an error message "Failed to open device generic/unspecified error" have you found a solution?