Experience with China Oscilloscopes?

[fchk]

Hi,

I'm thinking of buying a new Oscilloscope. I've got a Tektronix TDS220 now, but would like to upgrade on these reasons:
- 4 Channels
- more memory
- better screen

My options are:
- buying something better from Tektronix, LeCroy, Agilent,... used (I'd like to spend 5k, but I simply can't)
- go for some China equipment new

When I buy an TDS3014 or an TDS764, well, I can assume that there are certainly no serious design or software bugs in it and the only risks are worn-out crts or leaked capacitors. But I'm very unsure what to expect from Rigol or Uni-T or Owon etc etc.

Greetings from Germany
and sorry for the 4:0

Frank

[Time]

Dave has an extensive review of the Rigol DS1052E.  You might know what to expect after checking that out.

[squeezee]

The annoying thing about these chinese manufacturers is that while their 2-channel, entry level scopes offer very long memory depths; None of their more expensive ones do! (although compared to the tiny memory in the TDS220 anything will be an upgrade)

Exactly how much are you willing to spend? and i assume you're looking at something with the same 100MHz bandwidth?

[saturation]

Chinese oscilloscopes are a hot item on this forum.  Dave's Alexa ranking has been growing steadily, and Rigol and the 1052E are top search terms that lead folks here.

http://www.alexa.com/siteinfo/eevblog.com#

On reading Dave's original queries about the Rigol in 2009 elsewhere on the 'net [ where discussant dissect other Chinese scopes like Atten, Owon etc.,]  one scope that was left out but seems to perform very well, and has a company with a stellar reputation, is Instek.  Given a 4 channel scope has not yet been reviewed here, AFAIK, I'd like to hear what anyone has to say about Instek scopes and particularly its 4 channel models.

http://www.gwinstek.com.tw/en/product/productdetail.aspx?pid=3&mid=7&id=42

Face value, the Instek 1062 is a direct competitor to the eevblog popular Rigol 1052E, but most talk is about getting a low cost 100Mhz scope by hacking it, than examining all its functions as a general purpose scope.

http://www.tequipment.net/InstekGDS-1062A.html

Hacking aside and all things equal thereafter, the Instek offer better value with an upfront cost difference of $16.

[fchk]

The annoying thing about these chinese manufacturers is that while their 2-channel, entry level scopes offer very long memory depths; None of their more expensive ones do! (although compared to the tiny memory in the TDS220 anything will be an upgrade)

Exactly how much are you willing to spend? and i assume you're looking at something with the same 100MHz bandwidth?

I'm prepared to spend 2k Euros.

[squeezee]

Well the GW Instek 1062A is a fairly new scope, The old 1000 series were low-end 250MSa/s scopes with only 4KB of memory. The new 1000A series are 1GSa/s with 2MB of memory.

The 1062A is superior to the DS1052E in terms of specs, it has more memory (2MB vs 1MB), can sample at 1GSa/s in both channels and can use all 2MB in 1GSa/s mode. It also has a longer warranty and likely better software/support. If you aren't going to mod the Rigol it seems to be the better choice.

Unfortunately GW Instek or Rigol don't make 4ch models with the same large memory sizes.

For 100MHz, 4-channel scopes in your price range there's the Instek GDS-2104, Rigol DS1104B or Agilent DSO1014A. (the 200MHz versions are likely also within your budget)

The Instek is 1GSa/s with 25K of memory (not sure if that's per-channel), the Rigol is 2GSa/s with 8K/channel of memory, and the Agilent is 2GSa/s with 20K/ch of memory. (Agilent looks like a modified Rigol).

There's nothing 4-channel with a long 500Ksample memory until you get to the Tektronix DPO2014 or LeCroy 314A which are in the $3000 price range.

[saturation]

Anyone with hands on experience with the Insteks sometimes known by the moniker as the 'Instead of Tek = Instek?

http://www.memory-prime.com/

[alm]

If you want lots of memory, the used Teks (except the old high-end stuff, but that's still quite expensive and less nice to use) probably aren't the place to look, Tektronix always seemed to choose sample rate over memory (the TDS200 samples at 1GS/s, when the competition still did equivalent time sampling at 100MS/s, but has only 2.5kpoints memory). Even the current TDS3000C series is only 10k (is just TDS3k + some communication features like USB and software changes like Wavealert), you need to move up to the MSO/DPO series for more. I tend to use high bandwidth more than lots of memory, although ideally you want both. The memory is what lets you use that fast sample rate at anything but the fastest sweep speeds. I upgraded from the TDS220 to the TDS3052 for probably a little over $2k used (delivered to my home and plenty of time to test), it does 500MHz/5GS/s, but only 10kpoints memory.

Just yesterday I was troubleshooting a 50MHz pulse generator, and looking at the timing between internal 50MHz square(ish) waves and the amount of ringing. 100MHz would have made them all sinusoidal. There was a 4-40pF trimmer capacitor connected to the net I was probing, so a ~10pF loading from a passive probe would have a significant influence, and I used a FET probe (got it used, obviously). I also used the B trigger to get to the fourth pulse, and used advanced measurements and statistics (got the advanced analysis option via a hack). I wouldn't have been able to do this all with my old scope.

Features not found in my TDS220 that I use often:

• Extra bandwidth and memory (obviously).
• More measurements (eg. duty cycle). Statistics for measurements. Indicators which points it uses for measurements (let me see that it was measuring a pulse off-screen when I used the delay feature). Many of those come with the advanced analysis option.
• B trigger, delayed sweep/trigger. Lets my trigger on a certain event after the first trigger, possibly on another channel.
• Better UI for things like zoom (which is quite clumsy on the TDS220 in my opinion).
• Ability to set the inputs to 50ohms.
• Ability to use active probes.
• Variable persistence (DPO stuff).
• Much better screen, I like colors to distinguish traces.
• Ability to store four waveforms instead of two.

I'm not suggesting that you buy a TDS3000 series, just highlighting some (IMO) interesting features. One issue with Chinese manufacturers is accessories, support and documentation. Will they have software that supports the scopes for current OS versions ten years from now? Do they have good user, programming and calibration documentation? What about specialty probes? Even the standard passive probes might suck. The fastest passive probe from the Instek website, GTP-250A-2, has a fairly high input capacitance (17pF), and the specifications are quite limited. No mention of rise time, transient response and aberrations, and these are fairly hard to test. Companies like Probemaster have made some real junk, these might be no better. At least the Tek/Agilent/Lecroy people know how to design probes, and have the advantage of selling lots of probes (economics of scale). What if you want to use an active probe? Or differential? Or current? You can sometimes use those from a different brand, but most of the recent ones are designed to get power from the scope, so you need an expensive external power supply to use them with other brands. Plus most of the probes on Ebay seem to be from Tektronix or Agilent/HP.

Can't help you with the Chinese brands, the above reasons have kept me to Agilent/Lecroy/Tektronix. Agilent and Lecroy might have some interesting offerings, I've heard positive things about the current Lecroy scopes, but no idea about pricing. I just clicked on the GDS-1062A link on the website Saturation posted, all images and tabs appeared broken, not a good first impression . And the description about the 250MHz probe included 'Connect this sentence to the previous sentence.'

If I were to upgrade, I'd want more bandwidth in the process, not spending all that money for another 100MHz scope.

That's my $0.02 .

[DJPhil]

I thought the Instek might make for a fair fight in a review with the Rigol, so I tossed it up as a suggestion in the suggestion forum.

Just a heads up to let you guys know, and thanks to Saturation for pointing it out.

[cybergibbons]

My scope is a Instek GDS-1152A - I bought it over the Rigol because at the time, the Rigol hadn't built up such a following and wasn't so cheap.

I'm pretty happy with it really. Plus points:
* The sample rate is marginally faster when using both channels
* Twice as much memory depth (I find this very useful)
* Much nicer screen
* Less quirks/bugs
* Several firmware updates since I have had it
* Support requests have been answered quickly by e-mail
* Bought from reputable bricks and mortar retailer in UK - so covered by Sale of Goods Act.
* 150Mhz instead of 50Mhz - appreciable difference looking at pulses.
* USB interface has better coverage and appears to work better
* A set of controls for each channel - I don't like switching between them.
* Quieter fan.

Minus points (not essentially against the Rigol, just in general):
* More expensive
* No push buttons knobs, so resetting the position is a pain.
* No light up buttons, but then I have never seen the need
* Only has a 20Mhz LP filter on the channels, Rigol has configurable LP/HP/BP/BR. I've only quickly tried these on the Rigol and didn't find them to be very predictable (presuming Rigol does these in the digital domain and Instek in the analogue).
* Doesn't have slope triggering - again, after trying this on the Rigol I found it fairly hit or miss.
* Doesn't do +/- trigger like the Rigol, but then I can't see the need for eye diagrams at 100Mhz.
* Navigation through the 2Mpoint memory is a bit cack-handed. I frequently find myself adjusting the trigger point rather than just scrolling through the memory.
* Saving all 2Mpoint to SD card is very very slow (quicker with USB)

I think having a large amount of memory makes up for the lack of triggering options. I've used high end Tektronix before (there's even a user manual errata to my name!), and it is a different world, especially with triggering.

[alm]

* A set of controls for each channel - I don't like switching between them.

Doesn't really matter in my opinion. I've used both setups, both work fine, although switching between them is somewhat of a pain.

* Quieter fan.

This can be a real issue if you're trying to concentrate while the scope's turned on.

* Navigation through the 2Mpoint memory is a bit cack-handed. I frequently find myself adjusting the trigger point rather than just scrolling through the memory.

This issue gets more important as the memory depth increases. The navigation through the data on the TDS220 sucks, too, but with 2.5kpoints, you only have a few screens to go through, so it doesn't really matter.

I think having a large amount of memory makes up for the lack of triggering options. I've used high end Tektronix before (there's even a user manual errata to my name!), and it is a different world, especially with triggering.

Yep, lots of memory can to some degree make up for triggering, but it means you have more work to extract the data later. That strategy is used for capturing fast serial buses: don't trigger, just capture a gigapoint or so worth of data, and do clock recovery in software. I prefer good triggering, since it allows me to immediately get to the data I want, but both have their places. Advanced triggering is especially useful for rare events, if you want to capture an event that only occurs once a day, you can't do that without proper triggering.

[cybergibbons]

* Navigation through the 2Mpoint memory is a bit cack-handed. I frequently find myself adjusting the trigger point rather than just scrolling through the memory.

This issue gets more important as the memory depth increases. The navigation through the data on the TDS220 sucks, too, but with 2.5kpoints, you only have a few screens to go through, so it doesn't really matter.

My main issue with it is that there are two ways, once the scope has triggered, to navigate.

1. Zoom using the timebase and then adjust the position
2. Go the horizontal menu, press Window, use the position and timebase to but the window where you want, and press Window Zoom.

The problem with 1 is that you move the position of the trigger and the timebase, so you can't just press run again without setting it back.

2 is better, you can scroll around quicker. But it involves those extra button presses.

It does have a bookmark system, where you can mark points and jump back and forwards - useful for those times when you are looking for event x time after the trigger.

I agree with the triggering, there are times when nothing but a decent trigger will do.