Please suggest a good cheap USB 16CH logic analyser

[0xdeadbeef]

Yes I've seen both the devices called Jiankun and Kingst, maybe they are a re-seller, but I've even found links from the Kingst site to pictures of a Jiankun device so who knows?

I tend to believe that JianKun is the original name and Kingst a rebranding for the Western market. But indeed, that's up to speculation.

Yes, if you check their download link http://www.qdkingst.com/en/products, you can write your own protocol using their SDK.  I'm not sure about modifying existing protocols, but the SDK has a couple of examples for you to modify AKA Serial and SPI.

Well, an SDK is better than nothing, but personally I'm a big fan of using scripting languages for decoders.

As far as the memory goes, in the above post, I did identify the 2 ram chips which are Samsung K4T1G164QG chips and are 1Gbit each (sorry about the typo before) so that ties up with 2Gbit hardware memory (not 1Gbit like I said before.

Well, I can't say as I only found blurry pictures of the PCB, but actually the Kingst homepage also says 2GBit now (?).

I may be being dumb, but what's a specific clock or trigger input?  The analyser just reads specified data lines and the software has options to trigger on edges/high/low.

Purists would say that a LA without clock input isn't a LA but a protocol analyzer. With a clock input you can adjust the sample rate to that clock and thus guarantee that you store the levels at the correct moment.
I must admit though that I never really used the clock input of the DSLogic myself yet. Still nice to have one.

I took a look at the DSLogic Plus and thought it looked good.  At USD 149.00 I think it's good value (much better that Salea) although the specs of the LA5016 are better.  The DSLogic has a slightly lower max sample rate, at 400Sa/s and this is only with 4 Channels which, rather cheekily, they don't mention on their data sheet.  It's only got 256Mbit memory, so 1/8th of the Kingst.

The sample rate is somewhat meaningless without proper analog bandwidth and these probe cables add quite a bit of inductance at higher frequencies. The DSLogic at least has shielded cables.
About the memory: while I agree that having more can't hurt, I never ran into limitation with the DSLogic Plus. Usually sampling more than 1 or 2 seconds is just not convenient.
Besides, you can always switch to "stream mode" with lower sampling rates and sample up to 16G.

After saying the above, I think the DsLogic Plus would probably meet my needs as mostly my issues are with serial protocols such as RS232 and SPI so 2/3 lines are enough although I'd still go for the Kingst as the specs are better and hopefully more future proof. As mentioned above, I'm not sure about modifying Kingst existing protocols but that would be a good feature.

Nothing is really future proof nowadays. At least the DSLogic is Open Hardware and Open Source. AFAIK even the code for the FPGA is open source. So if everything fails, the community could take over. This being said, in contrast to the OpenBench Logic Sniffer, there doesn't really seem to be a community for the DSLogic and closing down the forum didn't really improve the chances that there will be one.

[picitup]

Well, an SDK is better than nothing, but personally I'm a big fan of using scripting languages for decoders.

I agree.  I think their SDK uses Microsoft Visual Studio to create DLLs and scripting is much easier and maintainable.

Well, I can't say as I only found blurry pictures of the PCB, but actually the Kingst homepage also says 2GBit now (?).

Well I've seen 2 x 1Gbit chips on the board, so 2Gbit it is 

The sample rate is somewhat meaningless without proper analog bandwidth and these probe cables add quite a bit of inductance at higher frequencies.

Yes I agree and the cables are long.  I'm making my own up, so will take your recommendation of using screened cables.

With a clock input you can adjust the sample rate to that clock and thus guarantee that you store the levels at the correct moment.

That's interesting thanks. I've learnt something

Nothing is really future proof nowadays. At least the DSLogic is Open Hardware and Open Source.

No, nothing is, but you want to get the best chance and better and faster hardware with more memory will help.  Open Source is definitely desirable.  I like PulseView, but the LA5016 is not supported yet.  The LA2016 is planned.

Yes, forums are very helpful and are always my first go-to.  It doesn't look like Kingst has one, which is a pity but as I mentioned, they reply to emails in a day so that's a +1.

Anyway after all this detail, I'm really pleased with my first analyser.  So much easier than a scope LA. I'm so glad I didn't fork out EU 500.00 for the Rigol decoders! I also ordered the LA2016 which is somewhere between here and China so that will go up onto eBay when it arrives.

Onward and upward..

Steve

[2N3055]

I also liked original Scanastudio, and had have bought ScanaQUAD because i liked that software and because ScanaQUAD had triggering.
Then I realized that it doesn't work with old software, and that new software is not written yet. That was 2  years ago, and in meantime, like you said, triggering was reduced to almost none compared to what was promised at the time of purchase, and software is so simplistic that it is sad.

I was active on forum, and some  user interface decision were made based on my input..

I basically gave up on them when Mr. Kamal decided to kill of forum to shut up negative feedback. Instead of developing their basic software to be at least as capable as old one they killed off, they spent two years developing wanky pen scope that nobody cares about....
Talking about going extra mile to ignore users and truth...

And then after 2-3 years, software is still in it's infancy...

I had high hopes for them, but they keep on doing things that are not in benefit of  users.
In order for me to respect  them again they would need  to :
- return multistage triggering
- make U/I that is more informative and uses was screen estate that is there
- write detailed documentation on object model for Scripting
- drop prices.

The way they are now, I wouldn't recommend their products to anyone...

[0xdeadbeef]

I also liked original Scanastudio, and had have bought ScanaQUAD because i liked that software and because ScanaQUAD had triggering.
Then I realized that it doesn't work with old software, and that new software is not written yet. That was 2  years ago, and in meantime, like you said, triggering was reduced to almost none compared to what was promised at the time of purchase, and software is so simplistic that it is sad.

I understand that the newest version of ScanaStudio is supposed to be scriptable. Didn't check it yet. And yes, as far as I recall, the improved triggering was partly done in V2 which stayed in Beta for years until it was replaced by V3 with more limited triggering. Still, to be fair, the triggering of most USB LAs is even more limited. The serial protocol triggering is somewhat unique (but again unfortunately not implemented in a generic way to actually use it for any protocol). Agreed, you can actually create some (!) complex triggers also in DSView but it's a GUI nightmare and you need to plan it on paper before even starting to program the trigger.

I was active on forum, and some  user interface decision were made based on my input..

The original forum was one of the reasons I bought the ScanaPlus.

I basically gave up on them when Mr. Kamal decided to kill of forum to shut up negative feedback. Instead of developing their basic software to be at least as capable as old one they killed off, they spent two years developing wanky pen scope that nobody cares about....
Talking about going extra mile to ignore users and truth...

Maybe it's too long ago but I can't remember a terrific fight there. Yeah, people were nagging about releases and the team kept up making promises and became silent when they were not kept. But nothing totally unusual.
Also before the forum was completely shut off, there was this abomination of a forum where you could ask questions which then were posted in one chaotic thread where you couldn't find it your own post half an hour later. About the same thing that NXP did with their forums a bit later.

And then after 2-3 years, software is still in it's infancy...

Again, I wouldn't say so and all I usually need is working in V3. I actually think the GUI is fast and intuitive compared to others. I'm a bit disappointed though that triggering is worse than in V2 and development is generally very slow. But my main gripe with my ScanaPlus is that I think its input circuitry is not optimal and due to its lack of (USB) data consistency checks, it can display total garbage signals now and then if it's overwhelmed by picked up noise.

Regarding slow development: the same is true for DsView. Actually, it's even worse. Some years ago, they introduced the dark design without any possibility to change colors. But I need a light background colors to use screenshots in reports.
I complained about that in the forum until it was closed down. Now, several years later, they finally added an option to switch back to light mode, but they did only exactly that. You still can't define your colors. All signals are gray by default and while you can change (only!) the signal colors manually, this change is not persistent. Hell, you can't even really store/restore a full setup in DSView. E.g. if you store a measurement and load it later, the "device" is changed, so you can't do any new measurements with this setup. Now this is missing basic functionality. And there is actually little to no progress in the DsView SW in the last five years. The only big improvement during its whole development was the speed-up of the Python protocol decoders which happened when the forum still existed. Since then, the development stalled.

I had high hopes for them, but they keep on doing things that are not in benefit of  users.
In order for me to respect  them again they would need  to :
- return multistage triggering
- make U/I that is more informative and uses was screen estate that is there
- write detailed documentation on object model for Scripting
- drop prices.
The way they are now, I wouldn't recommend their products to anyone...

I'm actually puzzled that they still exists. Few shops sell their products. Since they took so long to come up with the SP209, they didn't have a replacement for the ScanaPlus for several years. Instead, as you said, they developed that stupid scope pen thing that nobody ever wanted. And they wasted years developing that USB synchronizing absurdity for the ScanaQuad. It's not surprising that they had to raise the prices. They've just shut down their online store btw. Let's see if the SP209 can save them but I'm honestly skeptical.

[egonotto]

Hello,

if I make no error 0xdeadbeef has earlier  (few years) written that you can buy reduced ScanaPLUS. I bought one.

Now my ScanaPLUS is a ScanaPLUS V2.

The software is ScanaStudio 2.305.011

Are more ScanaPLUS now ScanaPLUS V2?

Best regard
egonotto

[0xdeadbeef]

ScanaStudio 2.x is a very old version which never came out of beta as far as I know. In my currently installed version 3.x of ScanaStudio (newest version is 4.x), my original ScanaPlus is correctly detected as such. If this is not the case in the old beta you're using, this is most probably a bug.
My understanding is that the only difference between ScanaPlus "V1" and V2 is firmware (FPGA code) which can't be updated for the end-user on the ScanaPlus (big design mistake IMHO).
The original ScanaPlus always samples all 9 channels internally. So if you only need e.g. two channels, the USB bandwidth for the other 7 channels is wasted. In the V2 model, they introduced the channel sampling selection that you just showed. The reason was that people complained about data inconsistencies for signals with higher frequencies (e.g. 10MHz clock signal). Since the ScanaPlus doesn't have any data consistency mechanism implemented (obviously to save bandwidth) at some point, there is a point where more data would be needed than available and the captured/displayed data becomes invalid. With the V2, you could increase the bandwidth by dropping channels. IMHO this concept was more of a panic reaction than a proper design choice. The SP209 has internal RAM now and is said to have a was of checking the consistency of USB data. But that took quite a few years and is still doesn't have USB3 AFAIK.

[ebclr]

This is inside LA5016 Jiankun one

[rsjsouza]

Thanks for sending this. Interesting how it says JK-LA-X016 but it is only applicable to the LA5016 and LA5032 (the LA2016 and LA1016 are clearly different as they use a smaller connector). Also interesting to see the sea of NC pins on the LA5016 - GNDs could make life easier for higher speeds.

For comparison, a LA2016 photo at Sigrok. 

[picitup]

My LA5016 is labelled Kingst and it has a slightly different layout with BGA ram and less components.  I've added a pic for all to see.

I've ordered a USB 1.1 hub to prove the memory really works and a 100MHz xtal osc and a couple of TC74VHC393FN dual 4 stage ripple counters, so will wire it up to produce 16 different signals (100Mhz, 50MHz, 25.....) and try all 16 channels with the USB 1.1 port at once.

Good point about the unused connections should be grounds.  I removed the top lid from my device and the PCB resides in a slot.  It really is stuck in tight and I can't get it out lol.

I'll report back when some progress is made.

Cheers

Steve

[picitup]

Well a bit of pushing and shoving and I got the PCB out by pulling the case apart slightly.  I think some bright spark may have put a drop of superglue on the PCB edge as now it's easier to get the board in and out.

When I put the board back, I couldn't get the case to line up.  That's because one edge of the case is male and the other female so you need to put the board in the right way.

I can confirm that all the NC pins really are NC so there's no option of 'improving' the device aka Rigol.  When I get a mo I'm going to short all the NC connections to ground so that will help.

The back of the board isn't riveting, but I've attached a pic anyway.

Cheers

Steve

[0xdeadbeef]

Seems that they optimized out the Max II CPLD in the newer (?) version. But also no visible signs of (automatic) length matching for the DDR2 RAM any more.

[ebclr]

Hi guess is 2 different things, 2 different companies, not the same, the price is also not the same

[0xdeadbeef]

Yeah, could be. Generally a problem with Chinese products. Even the original manufacturer usually doesn't care about clear model and revision numbering and several clones take over the case, the logo, maybe even the technical data, but you just never know what you're actually buying. Same is true to some degree also for the DSLogic. The original design changed twice or so but DreamSourceLab never cared to somehow document the official revisions. Plus there are several clones which look (nearly) identical but e.g. have cheaper probes or a different probe connectors.

[ebclr]

It's very common to Chinese copy the model name, The fact is that they "don't see English names " and they believe LT5016 is a coll drawing, look for phmeter PHS03, you will find more than 100 different product with the same model number, The fact is that Chinese only understand 这个狗屎

[0xdeadbeef]

I would think that any Chinese person will understand that characters in Latin script are letters. Obviously you can't expect that they always can actually read the letters or words.
At some point I wondered though if a trade name has the same meaning in China as in the Western world. I tend to think that at least for smaller companies, the focus is on creating a product and selling it with a good margin rather than on putting much interest in establishing a company name or being very proud of your products. In contrast to this, being proud to work for a company with a good reputation and being proud in your products might be more or a Western tradition (which is also going down the drain due to price pressure and marketing bullshit etc.).

[rsjsouza]

You guys made me curious. I opened my Kingst LA2016 and it is identical to the one at Sigrok.

[picitup]

Well the USB 1.1 hub arrived and to check it really does provide a bottleneck, I tested my USB stick with h2testw directly into the USB 3 port of my laptop and got 25MB/s write speed.  Then through the 1.1 hub I got 1.1MB/s so yes, it does provide a bottleneck lol.

The 100Mhz xtal osc has arrived, and I'm just waiting for the two ripple counters so I'll report back when those appear.

Cheers

Steve

[TK]

If the logic analyzer has internal capture RAM, then what USB transfer rate you have should be irrelevant.  If it offers real time streaming mode, then matters.

[picitup]

Yes, that's the test.  I want to see that the on board storage actually works.  Probably should get out more...

Steve

[picitup]

Well I'm still waiting for parts for the frequency divider.  The post is really slow here, I guess the donkey's ill again.

I woke early today and got to wondering what the accuracy of the LA5016 is and started looking at the PCB again.  There's only 1 crystal on board and that's the 24MHz for the Cypress USB chip.  I'm guessing the FPGA gets its clocks from the Cypress chip.  According page 26 of the the data sheet, the xtal pins of the 100 pin device are pins 10 and 11 (XTALIN and XTALOUT) but it is legal to drive XTALIN from an external 3.3v 24MHz clock.  The data sheet is here:

https://www.cypress.com/file/138911/download

The xtal on board has no markings, apart from 24.00 on the top so I'm guessing it's a cheapy one.  If required, it should be possible to fit a 24Mhz tcxo to improve the accuracy.  When I get a minute, I'll test the LA5016 as-is on my 10Mhz gpsdo.

Cheers

Steve

[TK]

It is my understanding that the 24MHz crystal is for the USB section of the Cypress chip.  Can it generate the high clock required by the FPGA for sample rates of 100-200MHz?

[picitup]

Yes it is, but there's a 20x PLL internally giving 480Mbps although I don't have any real idea atm.  I can't see another xtal on the board and the FPGA will need to be clocked.  I'll take a closer look....

[TK]

What about U14?

[picitup]

What's U14?  I'm outside my comfort zone here :-)

[picitup]

Oh yes, the chip 

I can't quite see the markings on my photo so I'll pop the case off and report back once I've drank more coffee :-)