Feel bad about my Siglent purchase

[Simon]

That makes sense being an accidental discovery, as the "allow to be hacked" business model would have been hard to sell without some prior evidence.

Rigol over-clocked the ADC by 2.5X, that must have produced hideous results 

What ADC did they use?

Best,

They were clearly new to the market and hit on this idea of end of line model programming as it was neat and easy and obviously never thought that someone would discover it so never locked it. I mean they were selling the exact same hardware, both 50MHz and 100MHz were 1GS/s.

I can't remember the details but when Dave did a tear down he looked up the part numbers and they were 40MHz but running obviously at 100MHz each as there were 10 interleaved to get the 1GHz sample rate.

[Bud]

Claims of such "business model" do not hold water. What do you think the manufacturing documentation would like? Would it have documents stating "We shall design with making hacks available ! ". Would engineering resources be assigned to design hackability into the process? Would a special Quality Assurance team exist to test such "features" ? This is laughable.

[bc888]

Sorry to read of your anger, but it seems that this would fall into the "Act in Haste, Repent in Leisure" category where you get to learn a lesson most of us have already learned (often the hard way as well).  It's a good lesson and thank you for sharing it. At least you didn't learn it with a car or a house where the ramifications could be super serious for you.

Take care.

[Mechatrommer]

That makes sense being an accidental discovery.. as the "allow to be hacked" business model would have been hard to sell without some prior evidence.

i have a feeling its not. after that, there's riglol out from no where.

Rigol over-clocked the ADC by 2.5X, that must have produced hideous results 
What ADC did they use?

you were not around were you? its enough to make hobbiests and people like Dave happy. what ADC used you can dig this forum. that day was the day like today when NanoVNA enter the VNA market. i bought $2K VNA from China about 2 years ago, and now we have $50 of the same from China too, should i feel bad? what do you think a VNA priced before i bought mine? before China make production? before that hobbiests can only dream. if talk about ethical, we should all buy Tektronix, Keysight, R&S, Lecroy or Anritsu.. did you follow some price trend of T&M equipments before and after China invasion? DMM? DSO? VNA? Trump did..

[Simon]

Claims of such "business model" do not hold water. What do you think the manufacturing documentation would like? Would it have documents stating "We shall design with making hacks available ! ". Would engineering resources be assigned to design hackability into the process? Would a special Quality Assurance team exist to test such "features" ? This is laughable.

Not really and I don't know all the details but I have owned 2 Rigol scopes and hacked both. The early one really looked like they just left a back door open that was a really convenient design feature that allowed them to make the same units and then decide which model to make them. The world discovered it and what I suspect was at first anger at the loss of revenue turned to surprise and realization of what they had unleashed as orders pilled in. There was at the time no option to upgrade through licence purchase as actually telling people that they had paid more money for the same hardware was probably thought to anger people.

My current model came years later. The concept of designing one model to cut manufacturing and stocking cots was established and so the official upgrade via licence code was introduced as it meant people may buy the base model and then be willing to pay more. It removed the need to hack for people who genuinely wanted to upgrade later.

But oddly enough getting access to the codes is very easy. i don't know how these were "leaked" or discovered but Rigol have made no attempt to get rid of websites that blatantly give away access to advanced features so it can be argued that they close an eye to it knowing that overall it generates more sales.

[Simon]

My basic understanding of ADC is that it's about being able to sample 0V and then 5V and then 0V with the sample hold circuit achieving the new voltage fast enough. The 100MHz band width is for a sine wave. so how much does the voltage actually change from any one sample to another that at most is 10% of the sine wave. As the frequency comes down the sample to sample voltage change is really small.

[mawyatt]

Claims of such "business model" do not hold water. What do you think the manufacturing documentation would like? Would it have documents stating "We shall design with making hacks available ! ". Would engineering resources be assigned to design hackability into the process? Would a special Quality Assurance team exist to test such "features" ? This is laughable.

Serious doubt any engineering resources were utilized to make the device hackable, just the contrary, you would need to spend resources to make it difficult to hack! So guess the laugh is on you 

Best,

[mawyatt]

Rigol over-clocked the ADC by 2.5X, that must have produced hideous results 
What ADC did they use?

you were not around were you? its enough to make hobbiests and people like Dave happy. what ADC used you can dig this forum.

I was not aware of Rigol until this year, or any of the these overseas type TE suppliers, as I've stated all the equipment we had and purchased was usually HP/Agilent/Keysight, Tektronix and R&S.

Finding a design that overclocked an ADC by 2.5X is just bad design practice for any product that sold to the public, unless there was some "special" feature of this ADC that nobody knew about that allowed this overclocking without disastrous results, why I asked what ADC this was! Every ADC I worked with or designed has had very bad behavior when clocking well past the design limit, from becoming non-monotonic, to missing bits, to locking up then periodically unlocking/relocking, massive non-linearities and so on, not something you want identified as "your" design in a product sold to the public, unless maybe it's a toy!!

that day was the day like today when NanoVNA enter the VNA market. i bought $2K VNA from China about 2 years ago, and now we have $50 of the same from China too, should i feel bad? what do you think a VNA priced before i bought mine? before China make production? before that hobbiests can only dream. if talk about ethical, we should all buy Tektronix, Keysight, R&S, Lecroy or Anritsu.. did you follow some price trend of T&M equipments before and after China invasion? DMM? DSO? VNA? Trump did..

You could have always purchased a used device and repaired/refurbished it like I did with a pair of Tektronix 2465 scope and a HP34401A and Agilent 34401A DMM. These didn't cost much and are still quite good instruments, even by todays standards. Agree most of the early TE stuff from China is in the "toy" category, and they have improved considerably, my 3 Siglent instruments are very nice and a great value.

A lot has to do with what you use the equipment for, a hobby or doing more serious work. If you were in the IC design sector like I was before retiring, a single IC design when you are pushing the State of The Art Performance can easily cost well over $10M for a test single fabrication run, then likely the TE will be Keysight, Tektronix, and R&S. A $100K DSO isn't much considering the cost of some IC developments, however if you are a hobbiest then the DSO will likely be much lower $ as will the DSO performance.   

Best,

[Mechatrommer]

Finding a design that overclocked an ADC by 2.5X is just bad design practice for any product that sold to the public, unless there was some "special" feature of this ADC that nobody knew about that allowed this overclocking without disastrous results, why I asked what ADC this was!

iirc its not 2.5X overclocked. there are 4x 200MSps ADC interleaved to get 1GSps, so each ADC clock is 250MSps, its only running 125%.. we cant see any bad signal shape out of it, either they are still within spec margin or Rigol knew some trick how to get around them. my DS1052E is still here doing fine, LCD is the one that is broken many months ago. i wont hesitate to buy used T&M if the price between new vs used is far off. i purchased a 15yrs old Lecroy DSO that otherwise will be priced 20X or more if new today.

[tautech]

I don't know how these were "leaked" or discovered but Rigol have made no attempt to get rid of websites that blatantly give away access to advanced features so it can be argued that they close an eye to it knowing that overall it generates more sales.

By tenacious code analysts present on this forum.

Their beavering away developed riglol and rigup and yet even more scripts recently for some series the Siglent range. There's lots of instrument investigations being pursued behind closed doors that the observant member can spot.

[LootMaster]

I put the blame straight on the foot of Siglent.

They missed the opportunity for an sds2354X-E @ 999 or 1100 $

Good honest transaction, will pay for MSO keygen and HW, no problem.

I needed THAT, but I'll take the luxury of new models. For an increment.

Caught me in the no man land. Pushed me to the edge. I don't care anymore, I'm like a rabid animal....I need MY GEAR.

[ataradov]

Dude, are you ok? You seem to blame everyone but yourself. This is not a good trait to have.

[LootMaster]

Awww not again.

The blame for what?

More features for me that I don't really need but will take to feel like a good nerd then move on.

More money for them to continue on and make new things... That I wont buy as I'll try to maintain this baby forever and ever like Mr Carlson's lab.

Hmmm... Maybe just a Power supply and DMM.

Change the caps every 5 years, maintain all buttons and blow it every year.

[mawyatt]

Finding a design that overclocked an ADC by 2.5X is just bad design practice for any product that sold to the public, unless there was some "special" feature of this ADC that nobody knew about that allowed this overclocking without disastrous results, why I asked what ADC this was!

iirc its not 2.5X overclocked. there are 4x 200MSps ADC interleaved to get 1GSps, so each ADC clock is 250MSps, its only running 125%.. we cant see any bad signal shape out of it, either they are still within spec margin or Rigol knew some trick how to get around them. my DS1052E is still here doing fine, LCD is the one that is broken many months ago. i wont hesitate to buy used T&M if the price between new vs used is far off. i purchased a 15yrs old Lecroy DSO that otherwise will be priced 20X or more if new today.

See #49 on previous page.

"We also forget so easily that those first Rigol scopes ran 40MHz ADC's at 100MHz by design to cut costs."


Some of the old LeCroy were rebadged Siglents,  like the Waverunner 2032.

Best,

[EEVblog]

i think it was an accidental model though that was discovered with the first Rigol scope that then became so popular that it set the scene for all that was to come in that market segment.

Yep, precisely this.

[rf-loop]

Finding a design that overclocked an ADC by 2.5X is just bad design practice for any product that sold to the public, unless there was some "special" feature of this ADC that nobody knew about that allowed this overclocking without disastrous results, why I asked what ADC this was!

This ADC, widely used by many cheap  scope manufacturers, is AD9288.
It is designed for 100MHz clock. All these versions, 40, 80 and 100MHz are same design, same silicon, same product. There is ONLY  one original AD9288 designed chip.

So, What are these 40MHz (80 and) and   100MHz ADC's.

Analog Devices have designed and manufactured ADC, product AD9288  what is designed for 100Msa/s.

But,  manufacturing process output is some amount of 9288 chips. Still all these are same chips.
But all chips do not meet every single specifications  what are set for 100MSa/s desing.
After they are made they need test. Some chips meet these specs and some not.
Do they throw away these what do not meet specs. No, they sell these.
So there is other specs, for 80Msa/s and for 40MSa/s
Chips what do not meed 100M specs may meet these others specs...  after then most of manufacture process output can sell.
They get -100, -80 and -40 marking.

Most of chips what have -40 mark  works normally in digital side just with 100M clock, even more. But they do not meet AnalogDevices specs for AD9288 -100.   They work but not meet specs.

If now someone set different specs what is enough for his own purposes  he can of course use -40 classified chips with higher freq if they meet his own needs. 

When Rigol or who ever use these they did not sell AD9288-100 chips with 9288-100 specs.  What they sell was and is oscilloscope with they own oscilloscope specs and what is inside oscilloscope is as it is...  they sell whole oscilloscope with this oscilloscope model description and specs  independent of what is inside it. They do not sell capacitors or named AD chips.

They can fill whole internal even with black epoxy. They sell oscilloscope with its specs and promise it meet given specs. This is what buyer buy and nothing else. It can include what ever except illegal forbidden things.



Not so seriously but...

I do not even know if it is soon more wise to make oscilloscope what is just permanently enclosed so that it can not even repair or open. Just if failure inside warranty time then just give new one and destroy failed. This can divide two or three parts, "black box" scope main unit and "black box" control panel and "black box" display unit. Three black boxes connected mechanically together and data bus between them. So no matter how it is done or what components they have used inside,  under clocked or over clocked.   

[mawyatt]

Finding a design that overclocked an ADC by 2.5X is just bad design practice for any product that sold to the public, unless there was some "special" feature of this ADC that nobody knew about that allowed this overclocking without disastrous results, why I asked what ADC this was!

This ADC, widely used by many cheap  scope manufacturers, is AD9288.
It is designed for 100MHz clock. All these versions, 40, 80 and 100MHz are same design, same silicon, same product. There is ONLY  one original AD9288 designed chip.

So, What are these 40MHz (80 and) and   100MHz ADC's.

Analog Devices have designed and manufactured ADC, product AD9288  what is designed for 100Msa/s.

This is what I was looking for, Rigol used the AD9288. That's an old tried and true pipeline CMOS ADC, and yes it is a fundamental 100MHz designed ADC. Today it's highly likely that this has been redesigned with a more modern CMOS process to reduce fabrication costs (smaller die). If you check the price on Digi-Key the 40, 80 and 100MHz versions are the same price which indicates a dominate yield at 100MHz (the die tests are fully automated, so basically all the working die are within 100MHz specs).

So they designed with the AD9288 100MHz ADC and used the 40MHz versions to save cost knowing that the 40MHz version maybe had some spec that wouldn't effect scope operation but wouldn't allow it to be branded by AD as a 100MHz version, so that's the "special feature" I was referring too.

BTW the AD9288 has a pin compatible 10 bit version, wonder if anyone has tried to upgrade the ADC with the 10 bit version? Of course this would be a lot of work, well beyond just replacing the ADC chip, since the firmware would need significant mods to deal with the 10bit ADC outputs.

Anyway, thanks for the identifying the AD9288 as the ADC used.

Best,

[Mechatrommer]

sibling threads...
https://www.eevblog.com/forum/testgear/feel-good-about-my-siglent-purchase/?topicseen
https://www.eevblog.com/forum/testgear/feel-medium-about-my-siglent-purchase/?topicseen