REVIEW - Hantek DSO5062B - A look at the DSO5000B series

[marmad]

[tinhead]

The "chinese toy look" is actually Tektronix fault.

You remember the TDS2000 series? As Tekway started to produce their first DSOs, they copied the Tektronix
enclosure and front panel - and they did it in typical chinese way - a "bit" cheaper. Not really critical, only the holes
around BNC and test signal out could be made better. As later the Tekway DST1xxxB (Hantek DSO5xxxB) was born,
instead of redesigning a bit the knobs/fron panel to made it a bit more modern, they just reused what already there.
Finally Hantek did the rest - they started to put the sticker probably by hand (this is what i think too, when i look on my
Tekway it is placed perfectly, when i look on all Hanteks i had in last 12 months - not a single with good aligned sticker).
Maybe not a big deal, but it would probably not costs that much per unit to make them as you said "more western designed".
Interessting that Tekway is still capable to place stickers properly, they even using different material for buttons
(they a bit softer, i ilke that more than these bit hard to push Hantek buttons).


The display is indeed optimized to use the scope above your eyes and not to put on desk, that was from the beginning
and can't be changed due mechanical design (you could flip display and flip then the data by changing 2 resistors, but
mechanicaly the display will not match into enclosure). For me this was always a big plus as all my lab instruments are
above or inline with my eye level, but sure for ppl prefering to have DSO on desk not the best solution.



Firmware, yeah .. i'm sure if you would use one of the firmwares developed by Tekway before they got bought by Hantek,
you would have to wait very long time for first crash. In principle (except the first 3 fw versions right after Tekway started
the production - which is typical for new products) i never ever got issues with the older firmwares.
Sure I did found some small bugs (mostly typos), but never saw my DSO crashing.
But that was long time ago as the firmware was 100% Tekway owned. I still remember the first firmware developed
by Hantek (they said "joined team from both manufacturers, but today i know that they in principle
fired everybody and started to do own things), i found immediately more than 20 bugs!
You can imagine how funny it was to write light years long email with bug description.

Sure today many things has been fixed, many new features implemented - but as you said - you can feel that the firmware
is still under development. One of the reasons might be that the management is more focused on new features/models
than on bugfixes for exiting models. They have now BM/BMV models (with funcy videohelp, 2gb flash, 2Mpoint meomory)
- which is really unique in this class of DSO (i meand video help and integrated storage).

They have also these new handheld DSOs 100% based on DSO5xxxxB (look for DSO1000B on Hantek website). These models
have already improved firmware, e.g. with LAN support (firmware and PC software imporved), even isolated models are available.
This is again something unique in that price class, high wfrm/s, 1GSs, LAN, isolated option etc. - no wonder that the engineers
have to be focused primary on these new models.

All we can do now is to wait until they ready with new models - all these improvements are anyway DSO5000B compatible,
so finlally we will win (same for SDK).



Timebase 2:4:8 hehe, look on Tektronix MDO, they changed "standard" too.
Now when you look on Hantek, the whole DSO is a mix of dividers:
- sampling (short mem) 1-2.5-5 up to 2us/div
- sampling (short mem) 2:4:8 for 800ns/div to 20ns/div
- sampling (short mem) 1-x-x (1GSs) for 8ns/div to 2ns/div
- sampling (long mem) 1-2.5-5 up to 20us/div
- sampling (long mem) 2:4:8 from 8us/div to 2ns/div
- timebase 2:4:8
- screen 16 or 20div
now try to find common divider for these values

Interessting that when i patch timebase table to 1:2:5 it does works too, so hard to say who got confused
and why they finally decided to use 2:4:8 timebase. It sucks a bit when you work with more than one
DSO (or have 6 fingers as i do), but after some time maybe not that tragic.



What you not mentioned in your review is the not existing dual cursor and tracking cursor. Sure the two standard cursors
are there, but i would be much better to have dual and tracking cursors too. They working on that,
got just message that there is beta fw with tracking cursor, see picture.

You spoke about the fast refresh rate, yeah this is definitely nice feature, however there are two others not mentioned
in your review - the DPO-like (or like Tekway said "wave-retentive technology") and the O.T. trigger.

There is also one useless feature implemented - probe check. In principle such function should be helpful to
set the probe compensation, but it was probably implemented by an idiot (or i expect too much).

With the long memory implementation, well this is a vice versa situation to what on Owon, Tekway probably
originaly focused on maximum available wfrm/s and these DPO-like features, but screwed up a bit the long memory
implementation. When in 1M the menu is responding much slower (seems that the engineer was playing too much with Tektronix),
sure still usable (somehow) but it could be much better.



Overall nice review! Thanks for that.

[saturation]

Most excellent review; I hope you do more.  You get to the salient points very quickly and with crystal clarity. 

You have the best video for a quick summary of the better low end scopes dissected at eevblog.

Owon was never on my radar as a scope of note, but you've certainly taken that view from zero to hero in under 15 min, with 3 short videos.

If anyone is looking for a scope, it should be mandatory to watch your Hantek review or the final video of the Owon review, as it has good overlap and puts together key issues, not just for these models, but for the qualities of the Chinese DSOs in general.

[FreeThinker]

Thank you marmad for the informative reviews. I am looking for a decent , cheap general purpose scope and you have done 90% of the hard work for me. The last 10% is of course up to me and mainly down to person choice. Think the Rigol will just edge it ( would be a clear winner with a owon display  ) but still time to change my mind.

[Nermash]

I've gone with the Owon, and so far I don't have any objections. FW can be improved, but compared to my previous dso this one has lightning fast UI and very low noise. I still would like to see waveform update rate comparison between DS1052E and SDS7102.

[rf-loop]

Exactly same signal with same connection and termination.
Hantek DSO5102B FW mod to DSO5202B, set: 4k, 1ch, normal capture (no average) 8s persistence so (it forgive littlebit perhaps most worse wobbles)
Owon normal not modified SDS7102VGA, set: 1k (10k give same result), normal capture(no average) infinite persistence and collected more than one minute before picture.

Signal is exactly same. Period 200us, pulse width around 10ns.

No, Owon picture is not cleaned. If I take picture with old very realiable Tektronix what can nearly trust, which one you think is more truth, hantek or Owon?  Most important thing in oscilloscope is signal guality this need always be first. What for are tens of features what you can use with produced garbage collection.

Look pictures, no need more explanation. This is not any "this I like more".  This is only simple data.
I can take more. This is only polite and nice picture  -- so Hantek does not lose face.

This difference is also natural.

Hantek use 8 parallel timeshifted 100MHz ADC's (interleaved). With 1GS/s they are clocked with 125MHz.

Owon use one chip fast speed (500MHz) double ADC chip (National ADC08D500 clone in this model). Inside this one chip is 2 ADC what can internally connect together for one 1GS/s sampler. (simple phaseinvert so also interleaved but... now clock signal is same!) with 2 channel use no any kind of interleaving, also ADC chip have internal self calibration procedure itself. So it is very clear that timing jitter situation is better (specially if look delta jitter between all 8 hantek ADC clocks) and also there is not balancing problems. 8 ADC in 4 chips can not be exactly indentical. There is 8 separate 125MHz clock signal and between these signals need be as less as possible timing errors. So in practice and in theory Hantek ADC system is light year behind Owon ADC.  Owon suffer littlebit 250MHz "noise" and this I think is small lack of good design. But then after look Hantek. Well, same problem but now 100 or 125MHz.
This ADC clock frequency change make many not so kind difficults for FW. Now  can also see that if you capture example 200ns/div. Then stop and look signal.. after you turn to 8ns/div (20ns ->8ns is freq change) whole captured signal disappear.

If look oscolloscope it really matter what is signal quality - if it is bad, all is bad. We can look if there are hongkong lights and tens of dingle dongles but oscilloscope is for analyzing unknown signal. If your signal is unknown you need trust that scope show it right enough. But yes, this is not interesting question, more interesting is trace color or how many different selections are in UI.

[tinhead]

No, Owon picture is not cleaned. If I take picture with old very realiable Tektronix what can nearly trust, which one you think is more truth, hantek or Owon?

simple answer - non of them (Owon, Tektronix, Hantek)

Owon is 100MHz model, most of HF is filtered out, i see less distortion which is typical for (almost) non-interleaved ADC.
When sampling with 1GSs the ADC works interleaved, but it happens internaly so clock phase is more accurate.

Hantek is 200MHz model, you see the HF affecting the waveform (when you set back to 100MHz and run both factory and self calibration you will see lees distortion). Sure, due 8 ADCs clocked externaly you will have more distortion, this is
typical for whatever model based on that design (many ADCs clocked by FPGA).

Tektronix - you can't compare analog scope to DSO due two factors:
- you will not see small jitter (of the source signal) on analog scope, the line will be only a bit brigther but exact this
jitter is creating a lot of distortion on a DSO
- let me gues, how much costs such 25yrs old analog Tektronix ? Exact the same price as a Owon or Hantek, right?
So don't compare like that, buy a cheap analog scope, feed same signal and you will see how crappy the picture is,
even on analog scope. We should stay below 1k $ when we compare something.


Now of course there are difference, even on expensive DSO you will see difference (and distortion) between
single vs. interleaved ADC design. Not that i like the clock on Hantek (if i would, i would not spend $$$ for clock
addon PCB on which i'm working now), but on the other side they at least using known manufacturers.
The ADC in Owon is a cheap piece of crap (believe me UNI-T would not replace them by x5 expensive org.
Analog Devices ADC without reason), org. National Semi ADC costs 200$, the cheap clone only 25$.


It is hard to say why so many manufacturers (Rigol, Instek, UNI-T, Tonghui, HanTekway, ATTEN, Siglent, others?)
are still using this kind of design (many small ADCs clocked by FPGA) instead of switching to more expensive
single or max dual ADC design clocked by dedicated chip. Knownledge? Can't be Instek GDS3000 is beautifuil design,
Cost factor? Cant' be, Rigol CA is cheap and using org. "single" National Semi ADCs.
Or maybe they just to lazy to change something and trying to earn the last possible $$$ with these designs ?
Anyway, sooner or later everybody who wish to stay on market will switch to less ADCs with dedicated clock.

Owon did it already, sure they cheatted a bit with these cheap ADCs, but this is already a step in right direction.
I would not hesitate and pay 100$ extra for an HanTekway with single ADC and 2Mpoint memory (shit, even with
SRAM instead of DDR2 like on Owon).

As i decided to buy Tekway there was no competitor on market, today there are some models where
the manufacturers tried at least to do somethign "similar" - they put larger disaply (ATTEN, Siglent, UNI-T) - but they
failed with fw implementation, others repalced ADCs (UNI-T) but still same kind of design (multiple ADCs+ FPGA clock),
others (Rigol CA) investing more money per unit for ADC, but forgot to replace display by more modern onea and
forgot to spend some memory, and finally Owon - as already said above - a step in right direction,
but knowing what inside they screwed totaly up the fw.

So now let's hope all these manufacturers will learn a bit from each other - Rigol CA with larger screen and
more memory would be good, HanTekway with better ADCs and DDR2 RAM would be good, Owon with
better firmware would be good , ATTEN/Siglent/UNI-T will be never good unless they understand that
cheatting with display, cheap ADCs, crapy firmware is what ppl expect from chinese manufacturers and
what the same ppl don't want to have in their lab.

Did i forgot someone ? Yeah, Instek .. they understood already, GDS3000 is nice DSO ... but
the price is already far away from all these models above.

[rf-loop]

Edi add: (forget)
yes you talk many wise words.
Also Chinese GW GDS3000 is nice (DPO) but this point is not nice if it is true:
"Independent Memory for Each Channel.  Record Length 25k points."

Well, what is samplerate with 500us/div.   IF 20div this whole mem, it is 10ms time to fill memory. Well it is 2.5Ms/s. (Owon cheap box 1Gs/s)
--------------------------------------------------



Here is now this same hantek modified back to just original DSO5102B
(yes cal after mod and all same as before pictures.)

And, this signal. It do not have any hf noise this amount what is meaningfull in this case.
It is old not so nice work horse HP8161A with Agilent 1999 improved  signal quality (I do not know how and what exactly)  improvement for mil lab use.

What I can not see with 350MHz Tektronix analog scope or 1GHz Tektronix analog scope what this cheap can see. No way!
I did not at all mean compare to these rolls-royces but... if I compare to these... and 2465 is enough good for tell that Owon is more close. Also I can look it with extremely old but also extremely flat response HP digital what BW is more like brickwall than gaussian. It is extremely level accurate from zero to 300MHz. (but sampling repetitive with 200Ms/s) also this show that Owon signal is good. (but this small hill after rise is related to Owon BW shape.)

All visible "noise" in signal is exactly hantek made and it is not coming outside by signal. I can look this signal with spectrum and this freq area is not in signal itself. (I trust my R&S and HP)

But Hantek have extremely more fast wfrms/s rate than Owon becouse Owon principle is classic DSO and Hantek is DPO (slow wfrms/s but still much more than Owon. And in this point need check which one select. If budget is limited around Owon - Hantek.
If need more samplerate specially lower than fastest t/div and littlebit better signal quality choose Owon if wfrms/s is not so critical or other things what are related to full speed memory.  Also if need external VGA, if need battery, owon. But if you need faster wfrms/s and want easy modify select Hantek, it have much more sometimes handy features in UI. Also full factory calibration after scope have been some time in use (aged) may give littlebit better signal noise.  Today Hantek FW is extremely buggy. But if you can live with these.. do not look these so much. And also Owon have some but far less.

Why UNI-T change ADC. Maybe UNI-T can not know how to use these ADC's. (do you know real reason for problems?) Maybe RuiFeng have some product quality problems after they start clone. Product version?
Original national have littlebit better signal quality. It can clearly read in datasheets. But how meanigfull these differencies are if other parts of oscilloscope make limits. Also Owon have small rule breaking in they design for this application with RuiFeng ADC's but not very bad. (some application notes they have not read (or care))

[marmad]

If look oscolloscope it really matter what is signal quality - if it is bad, all is bad. We can look if there are hongkong lights and tens of dingle dongles but oscilloscope is for analyzing unknown signal. If your signal is unknown you need trust that scope show it right enough. But yes, this is not interesting question, more interesting is trace color or how many different selections are in UI.

This seems a bit simplistic.

Yes, signal quality is important - but it's not everything.  An oscilloscope is basically an imperfect 100 year-old invention for creating an abstract 2D representation of what is essentially a 4D invisible event.  Yes, some scopes render a more 'accurate' representation of that event than others, but it doesn't make a bit of difference if the scope in question is allowing you to get your work done.  It's a tool - nothing more, nothing less.  I don't care if my hammer is missing a small piece from it's head as long as I can still pound nails with it.  It all depends on what you need the tool for.  I think that unless you're doing high-end work (which I would assume would require a high-end scope), small variations in signal representation are unlikely to affect most work.

Secondly, along the same lines, certain features that a DSO provide (or don't provide) can definitely affect your work as much as small variations in signal representation.  I think almost everyone has experienced buying a tool of one kind or another (hardware or software), and then not using it as much as originally planned - because of bad design, frustrating implementation of features, etc.  I don't care if my scope gives me the most perfect representation of a waveform ever seen - if it's a pain in the ass to use - or missing features that I need in order to use that representation adequately - I won't end up using it.

Lastly, there are times where signal quality is not the defining factor at all.  For example, what if I'm hunting glitches?  I'm looking for differences from the norm, so does the signal quality matter if it takes me 2 days to find an error which the scope is blind to?  I like the Owon scope, but once I discovered it's slow capture rate, I did many tests with glitch hunting - and they weren't very encouraging.  In one test I ran, the Owon captured a 1p/s glitch twice in 4 hours - in the same test, the Hantek captured it 72 times.  Which scope would save me more of my precious time?

[tinhead]

What I can not see with 350MHz Tektronix analog scope or 1GHz Tektronix analog scope what this cheap can see.

i see you didn't got what i tried to explain. The HP8161A have 20ps pp jitter, so far good enough.
Now when looking on that signal on 350MHz or 1GHz scope you will not see the 20ps jitter, maybe the line
will be 0.02mm broader but that's all.

But on interleaved DSO the situation change dramaticaly (single ADC based Owon will not get that jitter too, not even Agilent DSOX3xxx) every time one of the ADCs is sampling you have to add the 20ps, so finally you have 160ps jitter (worst case)
which is already visible on these DSOs.

These Hantek scopes can be calibrated only that good as good the calibration signal and clock source are.
Now let's hope they have high end signal generators for calibration (maybe they don't) also the clock
have already 100ps jitter (before FPGA) and FPGA itself 650ps jitter (with 20ps source clock)

Even with proper factory calibration you will have always some variations (200-400ps), add now the
estimated (when your HP8161A is calibrated) 160ps jitter and you already have what you observed
on screen - signal wobbling in all directions. Anyway, for a under 10k $ scope that's already good value (200-400ps),
and sure it could be better but not on multiple ADCs+FPGA design for under 1k USD.

But hey, your second picture from 100MHz Hantek looks already much better, so exactly what i said.

Also Chinese GW GDS3000 is nice (DPO) but this point is not nice if it is true:
"Independent Memory for Each Channel.  Record Length 25k points."

yes it is true, but i don't see a reason to cry about. 25k points is enough, just look on what this scope can
do (features, interfaces, modules) and not only pure memory. It does use proper ADCs, it is fast,
it is real DPO (well VPO), it does have VGA out, proper Trigger out, GPIB, SDK, protocol analyzer/trigger modules,
multiwindow firmware, search functions, puuh, everything you miss on Owon.

This is actually model which can compete with Agilent DSOX2xxx or Hameg HMO3/2xxx in many points,
the 10Mpoitns on Owon are nothing in compare (even it they nice).

For Owon fans, i can only hope the manufacturer rewrite the firmware from scratch, then these scopes will
have no competitors in under 1k USD class. Currently they just slow as hell, crappy firmware,
not working trigger out - but 10mpoints and manufacturer talking about "mature design".

I know you don't like some Hantek/Tekway firmware aspects (firmware updates vs. bugs), but they at least
doing something. Last Monday Tekway has been asked for tracking cursor - and 2 days later they implemented it
(that's why i edited my posting above), then the requester was still not happy with implementation and asked
for some extra things - and again 2 days later Tekway send new firmware
(no, please don't ask for it, this is internal use beta firmware only, with debug output so not for public eyes).

Just look on the attached picture and compare with the one i posted already today (my first post in this thread)
to see the changes.

Of course all these new features can also cause some bugs, but that's the price for new features
and i actually prefer to have manufacturer working with customer together instead of manufacturer
saying "mature firmware, no need to update".

[rf-loop]

In one test I ran, the Owon captured a 1ps glitch twice in 4 hours - in the same test, the Hantek captured it 72 times.  Which scope would save me more of my precious time?

Yes, as I tell it is Hantek strong point if compare to Owon with this test. Who need more wfrms/s than fast samplerate need better Hantek. But if you need take 10ms burst and look carefylly these pulses, you can not do nearly anything with hantek but Owon you can look it with 1Gs/s sampled. Both scopes have pros and cons.

Hantek may have around  300wfrms/s?

But this 1ps I do not believe at all - how you make it? What was real glitch time?

[Lightages]

He did not mean "one picosecond", he meant "one per second".

[FreeThinker]

A very good point marmad. This is why I have not (as yet) bought a digital scope. My main requirements are for signal tracing and timing and can be achieved with an analogue scope in the majority of cases. I would love a DSO but need to justify it to myself ('cos I'm really mean  ) and find that the only real advantage to me would be in glitch hunting and debugging (say) serial bus or similar. To this end I feel a Logic analyser  http://www.saleae.com/home/ would be more benefit so just what are the real benefits of a DSO?

[marmad]

He did not mean "one picosecond", he meant "one per second".

Yes - just edited that to clear up the confusion.

[tinhead]

In one test I ran, the Owon captured a 1p/s glitch twice in 4 hours - in the same test, the Hantek captured it 72 times. 

that's a difference, indeed. Do you remember in what timebase settings you did it? (i assume persistency was on?)

[marmad]

that's a difference, indeed. Do you remember in what timebase settings you did it? (i assume persistency was on?)

Since I had the scopes at different times, I jotted the notes somewhere when I first got the Hantek - I'll have to dig around and see if I saved the paperwork - I ran a lot of different tests.  I did run some tests with persistency, but I found it was easier to do long-term tests with pass/fail (it's a good tool to test the glitch-hunting abilities without having to monitor or use persistency).  It appears that both the Owon and the Hantek do very close to the same number of pass/fail tests per second (I think it's about 12 per second - someone else can verify) - but it seems the Owon is just capturing one waveform for each pass/fail test - while the Hantek is capturing many waveforms for each test.

[rf-loop]

Of course all these new features can also cause some bugs, but that's the price for new features
and i actually prefer to have manufacturer working with customer together instead of manufacturer
saying "mature firmware, no need to update".

First, I like both scopes, hantek and Owon, Im not fan for other, Im fan for Tektronix and Agilent or better say hewlett_packard. Both, H and O,  have weak points and strong points.
And as I have many times tell, Owon is DSO, hantek is DPO.

Yes this tracking cursor. I want it. This I have miss.

But then this Owon...
I have not hear any claim about "mature FW", who in Owon have tell this. This is maybe misunderstanding. maybe some in Owon have answer to someone but... you know these answers without any serious talking.  But also. This is product, look it, think if it pass to you and that's it. Next come new product and agen, look it and if it pass to you, buy. Why need think that I buy Trabant and then wait if factory update it to BMW.  So it is better that customer look product just in this time and look it with thinking if it pass to me just as it is and do not wait it is tomorrow different. If it is not now enough good do not buy.

But then if there are errors, these need repair to some acceptable level. This hantek have not done. They make new and new but not good repair work. Why? My hantek crash still least two times every time I use it. still after 4-5 FW updates. Also there have come more bugs. It is very frustrating example that scope change memory size without telling anything to user. . If I zoom over some point in time speed scale, I loose whole my captured signal. And same if I change speed, scope do not tell me that it have delete my selection with memory. Etc etc. If I shut off my scope I need take USB out. After starting put it agen and many small but frustrating things. But also many good things. In its price class it is good.

[tinhead]

My hantek crash still least two times every time I use it. still after 4-5 FW updates. Also there have come more bugs. It is very frustrating example that scope change memory size without telling anything to user. . If I zoom over some point in time speed scale, I loose whole my captured signal. And same if I change speed, scope do not tell me that it have delete my selection with memory.

yeah, you talking about bugs in 110728, 110801, 110806 and 110808 fw.
They are fixed in the version i have, so no more issues with memory and zoom.
I would normaly post it, but as said above this is debug version
(flooding uart and debug port) and not supposed for production units.

We should also be a bit more patient, HanTekway posted and send us so many fw versions during last months
(and some of them no supposed to be public) that ppl got a bit confused and sometimes pissed off about bugs.

Btw, all i can say about the latest beta version is that not only tracking cursor has been added, additionaly
the interpolation and filtering has been changed a bit, the signal looks a bit better now (less wobbling) but
it stocks from time to time a bit when i run it without debug connection (when i boot DSO without any connection).
When i connect debug port and uart it works good, but of course this is not supposed for default user.

[tinhead]

I did run some tests with persistency, but I found it was easier to do long-term tests with pass/fail (it's a good tool to test the glitch-hunting abilities without having to monitor or use persistency). 

yeah, make sense. When it was with persistency tested (i mean the 1ps) then the value could
match (just simplified 72/2 * 25 =900) as HanTekway is slower (like every scope) when persistency is on.
Interessting that these new HanTekway handheld DSOs/DMMs (DSO1000B) which are based
on DSO5000B/DST1000B (these numbers and names are really confusing) are also specified
with "Up to 2000 waveforms per second per channel (Normal acquisition mode, no measurement)".
I asked HanTekway to correct either website or user manuals, the mix of values (2500 or 2000)
is bullshit.

It appears that both the Owon and the Hantek do very close to the same number of pass/fail tests per second (I think it's about 12 per second - someone else can verify) - but it seems the Owon is just capturing one waveform for each pass/fail test - while the Hantek is capturing many waveforms for each test.

best case scenario - up to 140wfs per second can be captured while in pass/fail test, how many from them can be really
processed hard to say, i will have to setup proper environment to test it. Something like fail every second waveform
was showing exact the 50:50 in pass/fails counter, maybe truth maybe just coincidental.

[tinhead]

But then this Owon...
I have not hear any claim about "mature FW", who in Owon have tell this.

https://www.eevblog.com/forum/index.php?topic=4412.msg61013#msg61013

[colinbeeforth]

Hi marmad,

Thanks for the review.  I don't think I worry as much as you do about appearance.  We get very spoiled by the high standard of modern manufacturing.

I worry a lot more about the engineering, performance, signal integrity and ease of use.

There are very many aspects to DSO operation and performance, and few users exercise all aspects of a DSO except someone working in research and development of electronics.  When you push all the functions of the Hantek, it isn't perfect, but in terms of performance per dollar, it's much better than other scopes in the same price range.

I make a living fixing random pulse electronics (radiation detectors), and I've used a LeCroy 9310 for some years.  The Hantek is not very nice by comparison, but the highest sample rate is faster, and it does a lot of what I need for not a lot of cash.  For me, the Hantek DSO5102B is capable of doing a lot of the less demanding work I do, so it gets a tick from me.  The software bugs are really annoying, but power cycling is a quick solution if not very nice.

I think the Hantek/Tekway DSO, despite their buggy software, still gives much more performance than any other low cost DSO on the market.  Maybe  GW-Instek are equal or better performers, but I haven't got one to compare and they are not quite in the low price market area.

Good luck, Colin

[drieg]

I know you don't like some Hantek/Tekway firmware aspects (firmware updates vs. bugs), but they at least
doing something. Last Monday Tekway has been asked for tracking cursor - and 2 days later they implemented it
(that's why i edited my posting above), today morning Tekway has been asked to improve that tracking cursor
and few hours later they send new firmware (no, please don't ask for it, this is internal use beta firmware
only, with debug output so not for public eyes).

I can confirm what tinhead says. Well, actually the second improved version of tracking cursors took another 2 days (not only hours), but still, this is something you can hardly achieve with others (Rigol, Owon, ATTEN/Siglent...). First version they really implement in 2-3 days, then I asked them to improve it a little and sent them my proposal, then after 2 days I got the second version.
I wish they could fix other bugs as promptly as they did this

[marmad]

Thanks for the review.  I don't think I worry as much as you do about appearance.  We get very spoiled by the high standard of modern manufacturing.

It's not something I worry about, but it is the first thing you notice when you unpack something new.  Also, attention to detail and quality of manufacturing externally might reflect internal quality and attention to details in engineering and programming.  Certainly, the design and build quality of something could affect the ease of use and pleasure/frustration one has in using it.  That being said, I certainly wouldn't (and haven't) stopped myself from buying tools I needed, even though they were rather ugly or less-than-ideally manufactured.  I own quite a lot of Chinese tools 

I worry a lot more about the engineering, performance, signal integrity and ease of use.

Of course - I think I've made it clear in my reviews (and all that I've written in these forums) that it's a tool - and it's ability to provide the functionally YOU need (in regards to that tool) is the most important thing.

When you push all the functions of the Hantek, it isn't perfect, but in terms of performance per dollar, it's much better than other scopes in the same price range.

Well, you might be right - but to be fair, the Owon SDS scopes are slightly cheaper than the Hanteks - and provide some better hardware than the Hantek.  Again, it really depends on what you need to use the scope for much of the time.  For example, if you need an inexpensive Chinese DSO and you:
a) do lecturing - and need VGA out for projection;
b) do a lot of deep memory single-shot capturing and waveform inspection;
c) do occasional field work without access to AC power;
then the Hantek can't compete with the Owon.

[tinhead]

c) do occasional field work without access to AC power;
then the Hantek can't compete with the Owon.

well, there is new Hantek Handheld DSO-DMM, fully based on the DSO5000B series (but "only" 640x480 resolution),
so in principle you can get the fast wfrm rate and all the other DSO5000B features from battery powered Hantek,
but yeah this is not fair to compare handheld to desk device.

Well, actually the second improved version of tracking cursors took another 2 days (not only hours)

ups, i misunderstood your email, anyway 2 days is still really fast response. Sure, some bugs
could be fixed that fast (as e.g. the low filter issue in the beta fw with tracking cursor) but others
need some time, best example is the fine timebase bug while turning knob too fast - it happens not always and not
on every scope (and i replaced front panel PCB, replaced NAND/root fs/CPLD, sniffed for hours on bus to look for
a glitch causing this bug - and found nothing different between affected and working device).
The manufacturer fixed it finally then by rewriting the keyevent procedure.