Rigol MSO4000 and DS4000. Tests, bugs, firmware, questions, etc...

[hs3]

Three additional details: four channel only, smaller memory (14Mpts/channel) and slower update rate (60kwfm/s).

Could it be possible that they have done a lower cost re-design for the hardware with newer and cheaper as well as lower performance parts.

In the picture a few messages earlier it can be seen that there are 4 (could there maybe be 5th under that heatsink?) Xilinx Virtex 5 FPGAs and those are not the cheapest parts around.

I can't see in the photo what the exact part number is (would there exist an original better quality version of the photo that shows the part number on the Virtex?) but the cheapest Virtex 5 from Digikey is $250 for one piece:
http://www.digikey.com/product-detail/en/xilinx-inc/XC5VLX30-1FFG324C/122-1558-ND/1768318

Of course Rigol is not paying this kind prices but it's not cheap part in any case.

[Teneyes]

. I would find it hard they would save some money by simply removing the ADC channel interleaving of the MXT201's.

In the display ,pg3 ,shows 4chs and at 2GSa/s.  Perhaps dual DS4000 ADCs

Who is the target market?
Or dumping DS4000 end of life supplies?

[rsjsouza]

I can't see in the photo what the exact part number is (would there exist an original better quality version of the photo that shows the part number on the Virtex?) but the cheapest Virtex 5 from Digikey is $250 for one piece:
http://www.digikey.com/product-detail/en/xilinx-inc/XC5VLX30-1FFG324C/122-1558-ND/1768318

That is the exact Virtex P/N on my DS4014.

I imagine they redesigned the product with a smaller framebuffer (about ¹/₂ of the original ^wfm/_s) which may also require a smaller FPGA.

. I would find it hard they would save some money by simply removing the ADC channel interleaving of the MXT201's.

In the display ,pg3 ,shows 4chs and at 2GSa/s.  Perhaps dual DS4000 ADCs

Who is the target market?
Or dumping DS4000 end of life supplies?

I wonder if they missed the sales forecast for their DS4000 oscilloscopes and may be trying to recover some of the cost with a minimal redesign. I know many folks that wanted a DS2k with four channels, and that may be Rigol's response to that demand. 

[hs3]

I can't see in the photo what the exact part number is (would there exist an original better quality version of the photo that shows the part number on the Virtex?) but the cheapest Virtex 5 from Digikey is $250 for one piece:
http://www.digikey.com/product-detail/en/xilinx-inc/XC5VLX30-1FFG324C/122-1558-ND/1768318

That is the exact Virtex P/N on my DS4014.

I imagine they redesigned the product with a smaller framebuffer (about ¹/₂ of the original ^wfm/_s) which may also require a smaller FPGA.

Thank you for the closeup picture of the FPGA. Looks like they are using the larger FFG676 package with more IO.

So it would be this one for $287.50 each
http://www.digikey.com/product-detail/en/xilinx-inc/XC5VLX30-1FFG676C/122-1559-ND/1768319

If the DS4000 uses 5 pieces of these FPGAs that would be $1437.50 just for the FPGAs at the Digikey price. I imagine it should be possible to design a more cost optimized version these days.

Also if they are creating a cost optimized version that is intended to sell at lower price then unless they are planning to completely discontinue the current DS4000 they will probably need some way to differentiate from the regular DS4000. For that they may limit some of the capabilities even if there would be no real technical reason for that.

[rsjsouza]

Thank you for the closeup picture of the FPGA. Looks like they are using the larger FFG676 package with more IO.

So it would be this one for $287.50 each
http://www.digikey.com/product-detail/en/xilinx-inc/XC5VLX30-1FFG676C/122-1559-ND/1768319

If the DS4000 uses 5 pieces of these FPGAs that would be $1437.50 just for the FPGAs at the Digikey price. I imagine it should be possible to design a more cost optimized version these days.

Also if they are creating a cost optimized version that is intended to sell at lower price then unless they are planning to completely discontinue the current DS4000 they will probably need some way to differentiate from the regular DS4000. For that they may limit some of the capabilities even if there would be no real technical reason for that.

The distributors do not have budgetary price for 1ku for this FPGA, therefore it is very difficult to figure out exactly how much Rigol is paying for them. I would imagine they have quoted at least multiples of 10ku (2ku of 4ch oscilloscopes), where price would be significantly reduced.

One could figure out something from the retail price difference between a DS4012 and a DS4014 (around $300 at Rigol NA) - if Rigol makes 20% of GPM over cost (a low estimate) and the distributor gets something else (12~15%?), the remaining $200 would be split among two extra ADCs, two FPGAs, eight DDR and two SRAM devices.

Reducing the cost of the ADC would also be significant, given the budgetary 1ku price for the closest spec'd to the MXT2001 (dual 1GSPS 8-bit ADC) is the ADC08D1020 at $329.

(edited as I wasn't thinking straight in my parts evaluation)

[alank2]

If you translate the page to English it says "economical" model.  It would be interesting if rigolna carries this someday.

[hendorog]

Emona (Oz) just sent out a promo email about the DS4000E.
Here is the link:

http://www.emona.com.au/ds4000/

[John South]

Hi All - Just to let you all know Rigol have indeed released two new DS-4000E models.

These are released to fill a cost gap in the range between the DS-1000Z and existing DS-4000 for a 4 channel 100/200Mhz DSO. There are 2 models , DS-2014E 100Mhz 4 ch and DS-4024E 200Mhz 4 ch

Primary differences as has been pointed out are ;
Lower price - the above link has our local $ Aud pricing
Max sample rate 2GS/s vs 4GS/s
Memory depth 14Mpts vs 140Mpts
wfms/s 60,000 vs 110,000 wfms/s

If you follow the link as above to the Emona web site you can click on more info and download a 10 page brochure.

If you have any questions on these at all let me know.

[John South]

Sorry meant DS-4014E in the above post.

[rsjsouza]

Hi All - Just to let you all know Rigol have indeed released two new DS-4000E models.

These are released to fill a cost gap in the range between the DS-1000Z and existing DS-4000 for a 4 channel 100/200Mhz DSO. There are 2 models , DS-2014E 100Mhz 4 ch and DS-4024E 200Mhz 4 ch

Primary differences as has been pointed out are ;
Lower price - the above link has our local $ Aud pricing
Max sample rate 2GS/s vs 4GS/s
Memory depth 14Mpts vs 140Mpts
wfms/s 60,000 vs 110,000 wfms/s

If you follow the link as above to the Emona web site you can click on more info and download a 10 page brochure.

If you have any questions on these at all let me know.

John, thanks for confirming my initial suspicion: the DS4000E seems to address the demand for a "four channel" DS2000. A very smart move IMHO, given they only need to do a very minimal redesign.

BTW, do you know if the 2GSPS sample rate is active when all four channels are turned on? One of the pictures in the brochure seems to indicate that, but the DS4000 halves the sample rate every two channels.

[H.O]

So, comparing the 100MHz models you get 55% of the waveform update rate, 50% of the sample rate and 10% of the memory for 80% of the price?
Until someone buys one and figures out it's STILL the same hardware and unlocks it I don't think it's going to sell much.

[pascal_sweden]

It's actually weird that Rigol has introduced these new models as part of the DS4000 series, and not as part of the DS2000 series.

Before, the sample rate always defined the number in the DS series:
1 GS/s => DS1000 series
2 GS/s => DS2000 series
4 GS/s => DS4000 series

They could have used the following alternatives: DS2104A, DS2204A. Would be more logical to me!

[pascal_sweden]

If you check page 7 and page 9 in the Chinese product catalogue from Rigol,

https://www.google.no/?gws_rd=ssl#q=rigol.cn%2FFile%2FGuidAndView%2F%3F%3F%3F%3F_CN_201605.pdf

you can make the following observations:

When all 4 channels are used, DS4000 has 1 GS/s per channel, while DS4000E always has 2GS/s per channel.

When all 4 channels are used, DS4000 has 35 MPts per channel (140 MPts in total - shared for 4 channels), while the DS4000E series has 14MPts per channel.

Conclusion:

1) When all 4 channels are used, DS4000E has double the sample rate as the DS4000, but less than 50% of the memory depth.

2) When 2 channels are used, DS4000E has the same sample rate as the DS4000, but only 20% of the memory depth.

3) When 1 channel is used, DS4000E has half the sample rate as the DS4000, and only 10% of the memory depth.

So in terms of sample rate, the DS4000E series performs better than the DS4000 series when you use all 4 channels.

[TurboTom]

The information provided in the linked catalogue is not correct regarding at least my MSO4014 (and I assume all DS/MSO4000 models delivered so far as well).

Sorry - I guess it's a misinterpretation of the chinese text: The last figure (1GSa/s and 28MPts) refers to the digital channels of the MSO models which is correct. This means also the rest of my posting is more or less redundant since this information has been provided elsewhere before... Such is life 
_________________________

The model contains one 2GSa/s ADC for each analog input channel it has, i.e. the 40x2 models have two such ADCs and the 40x4 models four. Each channel pair has a processing engine (FPGA) and sampling memory (sufficient for 140MPts). If a single channel of a pair is used, the signal is fed to both ADCs and (apparently) the sampling clock interleaved, resulting in an output stream of 4GSa/s. If both channels of a pair are used, each ADC samples "its" analog input and the sampling memory is split between the two channels, resulting in 70MPts for each channel. A somewhat funny finding is that if on a four channel instrument three channels are turned on, the "single" channel of the pair where the other channel is unused, is forced to fall back to the 2GSa/s and 70MPts mode as well, maybe due to data display coherency reasons.

So it seems the "Economy" version of the 4000 series scopes lacks the sampling memory depth and the ability to interleave the ADCs to sample at twice their speed. Maybe this is a result of using smaller or less FPGAs. I doubt that the PCB and general hardware is that much different between the models. It's well possible the "Economy" models are just software crippled and can be hacked to full DS4000 performance (which would make them more than a bargain). I guess we can tell for sure only after the first teardown of an "Economy" model.

Cheers,
Thomas

[H.O]

And it's not correct with regards to the DS4000 either.
For single channel it's 4Gs/s and 140Mpts, for dual channel it's half (2Gs/s and 70Mpts). If you have a four channel scope then you have two dual channel scopes connected to one screen.

So, on a four channel version you can have 2 channels (1/3 or 2/4 at 4Gs/s and 140Mpts each) or 3 or 4 channels at 2Gs/s and 70Mpts each.

I don't read Chinese but I'm pretty sure that 1Gs/s you see splashed in there with all the Chinese tokens is reffering to the sample rate for the digital channels - which on the DSMSO4000 is 1GHz.

[rsjsouza]

And it's not correct with regards to the DS4000 either.
For single channel it's 4Gs/s and 140Mpts, for dual channel it's half (2Gs/s and 70Mpts). If you have a four channel scope then you have two dual channel scopes connected to one screen.

So, on a four channel version you can have 2 channels (1/3 or 2/4 at 4Gs/s and 140Mpts each) or 3 or 4 channels at 2Gs/s and 70Mpts each.

I don't read Chinese but I'm pretty sure that 1Gs/s you see splashed in there with all the Chinese tokens is reffering to the sample rate for the digital channels - which on the DSMSO4000 is 1GHz.

I can confirm what H.O. said.

[Teneyes]

So the DS4000E is locked at 4chan and 2GSa/s, or not needed for 200MHZ all channels

Note from Emona

Hi  - I'll check with Rigol and wether or not there are independent ADC's per ch. or shared and let you know.

I don't yet have a unit in country to try on , our demo stock are on order but still a few weeks off.

[CustomEngineerer]

It's actually weird that Rigol has introduced these new models as part of the DS4000 series, and not as part of the DS2000 series.

Before, the sample rate always defined the number in the DS series:
1 GS/s => DS1000 series
2 GS/s => DS2000 series
4 GS/s => DS4000 series

They could have used the following alternatives: DS2104A, DS2204A. Would be more logical to me!

Completely agree. And also if the 14Mpts/channel is correct, 14Mpts x 4 = 56Mpts, same as the DS2000A, with memory option enabled.

[pascal_sweden]

Moreover, if they would have introduced it in the DS2000 series, they could still have used an A for Advanced, instead of an E for Economy.

By definition, it is not good to use a letter in the model that indicates it is the Economy version
This way all your test & measurement friends know that you selected the model for Economy reasons.

I remember that some computer brands in the 90's even used the letters CR in the model number,
to indicate Cost Reduced. That's even worse

Anyhow, it should have been introduced in the DS2000 series for the reasons that I stated before.
(Sample rate = 2 GS/s => DS2000 series)

[TurboTom]

Since it's not possible to offer a four channel scope in the DS/MSO 2000 form factor, and the market was demanding such a machine with more than 1GSa/s distributed over all four channels, it's just logical to follow the road of least effort and re-use what's already available. An instrument with an enclosure of the DS4000 series with a DS2000 designation would appear somewhat out-of-place. If it was a wise decision to introduce an "Economy" version is another question. It seems Rigol followed that road with several new products (see the DSA832E). Maybe a new designation for the oscilloscope (like DS3000 series) would have been the better coice.

In the end, designation doesn't matter as long as it performs well, it's reasonably priced and it's hackable... 

Cheers,
Thomas

[H.O]

I'm with Thomas. 2000, 4000, A, E whatever. Who cares about numbers on a sticker?
But again, I don't know about this "economy model"....the economy in it really doesn't look that good to me, the price difference is too small IMO - IF it doesn't turn out to be the same hardware of course.

[Teneyes]

, it's reasonably priced and it's hackable...

LOL
Upgradable ? NOT easy, I think, 
Could just be the DS2000 56M memory, different front end filter Amp.

[pascal_sweden]

Another alternative could have been to use the form factor of the DS1054Z, and call it DS2104Z

[Teneyes]

FYI , for USA , NA
---------------------------------------------
Hello
As per our chat last night.
We are currently accepting pre-orders for the DS4014E.
The estimated arrival for the US market is around 8/15.
Let us know.

Thank You,  Stephen
TEquipment.NET        205 Westwood Avenue | Long Branch, NJ 07740
Email: SalesTeam@TEquipment.NET

[rsjsouza]

, it's reasonably priced and it's hackable...

LOL
Upgradable ? NOT easy, I think, 
Could just be the DS2000 56M memory, different front end filter Amp.

The cost reduction has to come from somewhere, and there is plenty to reduce in a DS4000 as I stated before (FPGAs, ADCs, memory). I suspect memory and GSPS are not upgradeable, but perhaps decoders and other functions. Obviously we would only be sure after opening one up and comparing IC markings.

It's actually weird that Rigol has introduced these new models as part of the DS4000 series, and not as part of the DS2000 series.

Before, the sample rate always defined the number in the DS series:
1 GS/s => DS1000 series
2 GS/s => DS2000 series
4 GS/s => DS4000 series

Apart from the fact that:
5 GSPS => DS6000 series

Therefore I suspect Rigol is naming their families after "look and feel" instead of sample rate - don't forget these guys also have a larger screen, different button arrangement, etc.