500 MHz 8b 2GS/s oscilloscope

[petar]

Hi folks.
My name is Petar (that would be Peter if I were Englishman).
I am former air force pilot.
I am also electronic engineer.

About eight years ago I finished designing of a PCI data acquisition card.
It took me almost 3 years, but I designed everything from the starch including PCI bridge (scatter-gather DMA…), SDRAM controller, I2C, SPI, oscilloscope core and many others in VHDL. I was using Spartan2. The card was two channel oscilloscope and arbitrary waveform generator. Analog front bandwidth was about 200 MHz but was slow 50 MS/s 12b ADC and 100 MS/s 12b DAC. It is blazing fast and one has opinion that it is analog scope. I wanted to make a firm, but circumstances made me to give up (waiting for a better time… maybe newer).
Time elapse and I was doing lot of things… ASICs design in verilog…, DSP in FPGA (virtex, virtex2,4,5,6,7)… (software defined radio, spectrum analyzers…). I was the first one that used new National GHz ADCs (long time ago) and had a very bad experience with National engineers and find the bug for them...
Everything that I was doing with a collaboration with “big” firms makes me to believe that there is a few people that really knows their job and a mass of engineers that worth nothing (regardless of their engineering degree). Almost always they put the finger on me and told that I don’t know…. and after tons of argues and lot of time the truth came to the surface and I was right… There is a lot of firm that I designed parts of their chips… or parts of IPs… Of course they don’t know me (they know men that I was working for, and all of them hide me from direct contact with actual customers)…
I know PCB, RF, FPGA, ASIC, DSP. For many fields I am expert and for many I am good at, but not the expert. One would say that isn’t possible, but one is wrong.
Forgive me if I am boring.

Let’s talk about my current steps.
I am designing oscilloscope that would have two channels, 500 MHz bandwidth, 2GS/s 8 bit ADC per channel, 250 MS per channel memory depth, analog or digital or external trigger, normal oscilloscope offsets controls, 400V DC+AC protection at inputs, two analog output channels with 250 MS/s 14b DACs, 20 digital inputs. Because I couldn’t find cheap enough and small display I decided to make an oscilloscope like “picoscope” but not the PC based. It would have VGA out for normal PC monitor and PS2 for mouse.
It will be blazing fast because I decided not to use windows or linux for the job but to generate the picture direct from FPGA and to monitor the mouse (that is only thing that I currently don’t have FPGA IP prepared but I made calculation and strategy for the job… and sized one of the FPGAs for the task according to my calculation). It will have also USB2 slave port (I made a USB with cypress chips long time ago and I got about 26 MB/s to and from PC).
I finished the RF simulations of analog front end and schematic.
Currently I am designing layout.

My desire is to establish a firm.

I am poor man from “west” perspective, so I have only enough money to make couple of oscilloscopes.
I know where the money “lives” (in bank) and where I could borrow it, but I don’t want to do that because those that have money want all for them…
I am asking if it would be possible all of us that’s want that scope to make an “organization” or some “community” that will collect money in advance and make a chip oscilloscope that cost at least 4 times less then when you buy it from Tektronix or other “big” firms.
Believe me or not some parts that all us buying from Digikey can cost twelve (or higher) times less when ordered in large quantity direct from suppliers and that’s exactly the most priced parts. Of course I can make couple of scope before that and somehow presents to community…

I am asking all of you, when I finished the scope what would make you believe me.

My previous life only tough me that bombs drops on my head from “west” and I personally think that what I want is utopian, but why knows maybe human kind will be better and could recognize the truth.

PS. One of my colleague has firm at US and when he begin to establish the firm when pricing product (software) started from about 600 $ and had a few of customers and have been told that if the price would be about 1000 $ that would have more customers. When he rise the price to 1000 $ he couldn’t help himself from lot of customers. The software that he developed was the same… But no one believed that it worth anything when price was low. How men are weird.

Thanks,

[EEVblog]

Believe me or not some parts that all us buying from Digikey can cost twelve (or higher) times less when ordered in large quantity direct from suppliers and that’s exactly the most priced parts. Of course I can make couple of scope before that and somehow presents to community…

FYI
Digikey also sell medium to large qty at fairly close to the manufacturers direct price. And you can get "DigiReels", to order just the required number, so no wastage.

Dave.

[EEVblog]

I am asking if it would be possible all of us that’s want that scope to make an “organization” or some “community” that will collect money in advance and make a chip oscilloscope that cost at least 4 times less then when you buy it from Tektronix or other “big” firms.

It's called KickStarter
www.kickstarter.com
You have to be US based though.

Dave.

[petar]

Hi Dave,
Believe me or not. I am in collaboration with some big firm and I don’t want to name it because of some reason…
12 times is true.
I am not talking about resistor or capacitors…

Thanks,
Petar

[petar]

Dave,

Thanks for the link.
It is good and valuable.
I read it couple of minutes and find that what I am doing is slightly different.
It says: Creators aren’t expected to develop their project without necessary funds, and it allows anyone to test concepts without risk.
But I spend entire decade of my spare time to collect knowledge and make IP necessary for the project and I don’t want others to fund the developing I already did it.

Thanks,
Petar

[EEVblog]

Hi Dave,
Believe me or not. I am in collaboration with some big firm and I don’t want to name it because of some reason…
12 times is true.
I am not talking about resistor or capacitors…

Neither am I.
Look at say the single reel prices for most semiconductors from both Digikey and direct from the manufacturers and you'll find the Digkey prices match almost precisely. So if you can get the parts cheaper then they are likely form the gray market, or you have a special deal with some semi maker that usually only the big guys get when buying 100,000 units or something.

Dave.

[mikeselectricstuff]

It does depend a lot on the manufacturer, but if you're buying full reels, and can live with some leadtime, other franchised distributors are often cheaper than Digikey, but rarely more than a few tens of percent lower.   
There are some oddities, e.g. FPGAs whch can easily be half the Digikey price once you talk to a specialist distributor, even as low as 100x

[petar]

Here are some picture of my previous card I was talking about:

[petar]

The relays that could be seen on analog portion are very small 6mm reed contactors and I used omron relays with it magnetic field to activate the contactor underneath it.

We both know that semiconductor worth nothing. That is the sand, that each river carry a lot of…
We “all” think that some multi billion transistor processor worth let’s say 20$ in production, no that’s the lie. It worth nothing.
Someone would say it is true why you don’t make it and…
But I am certain if one from “poor” country would make such a thing, the guys from west would kill him, or kill entire country whichever is easier to do.
I am also certain that engineers that did all the jobs are less than 1% of the price we are paying for the product.

I saw that you say that: “If someone from China wants to copy your product you could do nothing”. Bad China… No, we are all humans we are all same. China is under total west control. All that is “steal” from the west is under direct control of the west people, so lets say this way: “If one from your neighborhood wants to steel something from you than he goes to China to make it there…”

In human history there is no single event where someone that has power to kill, and knows that he would win the war waits until other side becomes stronger than him. Every time when someone is shore that he would win the war was inevitable. Why USA don’t attack China (the only answer would be: China is under total control.)

[aluck]

Petar, take a look at those guys: http://www.seeedstudio.com/depot/about_us.html

[Spiro]

So, what to you want to say with this story? Do you want to create open
hardware scope but you want that nobody to copy you design? You want
to earn your first 1M$? I know. You want to ruin all big names as revenge!!!
Hahahahahah... tragic. 

[kaz911]

Petar - I have quite a lot of experience on "the other side" as either investor or mentor for start-up companies. If you need any hints I'll be happy to help. I have considered setting up a "Crowd sourcing" platform for B2B products (Where Kickstarter is more a B2C oriented platform)

Unfortunately products directed at the EE market are not easy to get finance for because a lot of Business Angles and Investors does not understand the products. And then many Engineers are not versed in business planning - so their budgets are "full of holes" so the investors get scared away.

And everything is more expensive than one thinks. When you design products and build from scratch - a normal rule of thumb is that if you product cost $100 to manufacture - the selling price should be at least $300-$400 to sustain the rest of the "company". Many companies archive that by having a range of models like for scopes - where you earn a little on the small units - but a lot on the high end units - so average return on each item sold is 3 - 4 times production cost.

But that is based on a distribution network - with own sales/support & R&D.

[NiHaoMike]

Better update the board to PCI-E or USB 3.0 before trying to "sell" it, you won't get much demand for something that only works in some old server boards.

[petar]

I saw lot of double crosses in my engineering career.
What I have leaned from my previous life is that humans generally doesn’t recognized real value.
Talking about investors, once a firm from west that had a nice idea for cell phones earn money from “venture capital” and doing lot of designing until was time for investor to see first results, but they didn’t know the job and they didn’t got anything working to show. That firm engage my current west firm that I working for, and we made almost finished product based on Altera FPGA for them in about 3 months. They had what they want and had a presentation for investors on time, and they of course didn’t tell investors that someone else did the job for them. Of course they gain lot of more money from investors and wanted to make an ASIC for the job. We can design ASICs, but they left us and engage another firm to do ASIC for their job.

This tells the fact that actual investors didn’t have any knowledge and that he was blind…

I don’t see the difference between that investor and all of you.

My question was clear, do each single one of you guys believe me or not and what single man can do to others that others starts believe him that he telling the truth?

I don’t want to start open hardware at this time.
I would certainly like to ruin big names, they are only big in our minds and they makes our lives worst.
If the price of hardware would be so small then no body would have a need to make a DIY and open things…

That are the pictures of my previous card, current one would be stand alone (not PC based).

Please all of you forgive me for my bad English, I know I make mistakes all the time, but I am certain that you understand me and that’s what is important.

[EEVblog]

My question was clear, do each single one of you guys believe me or not and what single man can do to others that others starts believe him that he telling the truth?

From your first post, I'd have to say no, it was just talk, like countless people have done before.
But you have posted photos of some excellent looking hardware, so I'm sure most people would believe you are most capable of doing this now, at least on the hardware side.

I don’t want to start open hardware at this time.

Well, that might be a problem.
Outside of a KickStarter type project, I think realistically it's not going to be easy to convince people to invest in your project, because for starters this hardware will not be cheap in small to medium volume.

The way these things usually work is that you produce a prototype of your final hardware first, and make it look as good as possible and as close to the final product as possible. Then you show it off publicly and maybe take pre-orders to get seed money that way for a first batch.
That's basically how kickstarter works too, but you don't have to use Kickstarter, you can just have your own website and a PayPal button.
But really, you need the final hardware to show off, because no one is going to invest and then wait 12 months for you to get something out the door, they need something tangible to lust after. That means you need to show the final (or close to it) hardware working, measuring signals, and some characterisation of it's performance etc.

Dave.

[EEVblog]

BTW, this Australian company are doing something similar:
http://www.screenscopetraces.com/index.html
but not in the same performance category.

Dave.

[BravoV]

I'm interested, what is the approx cost for this cool 2 ch 500 Mhz scope ? $100 ? $1000 ? or else ? At least you can give us the ballpark number as a good start since you claimed this will crushed big companies like Agilent, LeCroy and Tektronix right ?

Oh yeah, btw, which country you're currently living at ? Anywhere near or in Nigeria ? Honestly, this is just a random pick since you never give us any clue about your self nor any details about the scope.

I'm really sorry as English is not my native language too, I'm asking this because the current details I've collected from this thread are only some faint infos that you were a fighter pilot, an EE, some board pics and your political view on China vs Western world, correct me if I'm wrong.

[gregariz]

As Dave more or less alluded to.. I think you are going to have to finish the product first.

If you cant afford to finish it you could set up something like a kickstarter or your own website. Nothing to lose and it need not be an open source project. Otherwise you could try to bootstrap yourself by selling a few at a time until you have the cash for higher volumes.

Just from personal experience I think you should have a close look at what design decisions you make. Doing everything from scratch yourself in an FPGA may seem impressive from an engineering standpoint but usually in a commercial environment it means that development time is excessively long and the FPGA sizes you will require are going to be somewhat expensive. For example, do you have any idea what the cost differential is between developing the VGA interface versus simply using the PC display? Is it worth the effort in time and money? Who are your customers? how will you sell to them and what do you think they will pay?

Finally I know that anyone who has spent more than about 5 years in an engineering role becomes extremely cynical (and for good reason) but there is no grand conspiracy on this board to somehow stop you from doing your project. No-one is interested in your politics. Good luck.

[robrenz]

Big companies have to pay their engineers, salesmen, and administration etc..  Yes they make lots of money but they also have to spend a lot.  If you charged $35.00 an hour for ALL the time and expenses you have invested in this so far and the additional hours and $ to get it to a sellable product and had 100 units made in Taiwan your scope would probably cost more that a Tek or Agilent equivalent.  And who is going to pay you for the tech support and warranty service?

I understand a labor of love as many of my personal projects are just that. I don't get compensated for the countless hours I spend developing a product and I don't expect to.  My pay is the enjoyment of the creative process. 

IMO your thoughts that you can produce and support a competitive scope in small quantities for substantially less than the big guys is unrealistic.

[petar]

BravoV,

Why are you so upset about all the details you want to know?
Do you own those "big" companies or work for?

I didn't tell that I will crushed big companies like Agilent, LeCroy and Tektronix.
I told that I would certainly like to do it.
That's big difference.
I just don't have power to do it. My customers will be my power and any honest man on the planet will be my power.

I am European. I speak English and Je parle Francaise. I have two university degrees. I am white like snow. My mother was Croat and my father is Serb. I am from country where once live Nikola Tesla, have you heard anything about him?

First of all I am designing the oscilloscope for myself (because I personally believe to myself and have no fear of double crossing myself) and I know that I couldn't buy such a quality and features for the price of single unit BOM at Digikey. I will finished my scope and no single person in the world have to by it (I would certainly like that people buy it, but I will not kill myself if nobody want it). I love electronic, that is the reason of doing all that I do. The second reason is that I know that my work is state of the art and I want to have state of the art equipment in my lab (not just a scope and not just a single one).

Second, I had that PCI card long before let say picoscope establish itself as a firm and I see that picoscope is now considered as respected firm and I see myself still as a poor man (relative poor).
Long time ago I search picoscope site for the information on input analog offset adjustment range, and I had been searching it almost 3 hours. I couldn't believe my eyes that picoscope don't have offset adjustement...

On the other hand I hate managers that ruin the product performance just by using low quality parts and the quality drops let say many time for a couple of $. Also Chinese oscilloscopes are low signal quality and full of bugs (at least what I saw).

As I am working with such a equipment every day one couldn't believe how many bags have equipment that cost let say 20k$. Shame, blame, "big companies", all garbage.

Price of oscilloscope is a function of BOM (in my case) and it depends lot of on quantity.
Schematic is optimal for good results and there is no single chip that isn't necessary.
I will see (ASAP) on quantity of one what would be the cost of BOM at Digikey (it is certainly not 100$).

Can you tell me what is the price of 500 MHz oscilloscope, 2GS/s per channel, and 250 MS per channel memory depth? (100 $ or what?)

Thanks,
Petar

Gregariz, you are right and many of my colleagues told me the same thing, but if I use some processor and all the stuff around it SDRAM....., and of course must pay if I use windows... to display... than it would cost me almost about 100$ per scope, and I already have FPGA for the job. Certainly your approach would be faster development time. Oscilloscope screen isn't movie (it is big darkness-nothing  and very few something) when I compress that big nothing it is so little (few MB) of data....
The compression wouldn't been MPEG2, no, it would be "my" compression which is so simple... again, it is not film nor picture of some nude woman. Only thing I need to do is just one time to decompress and resize the picture (to adjust to  monitor size) and to change so few pixels on the fly while oscilloscope works.
And all of that would cost me about 100k gates and few DSP48 (oh mistake few multipliers) in Spartan3AN and extern SDRAM.

Thank you all I appreciates your suggestion.

[petar]

Hello Robrenz,

I agree with you, lot of hungry in the chain…
But when speaking of support… when equipment works no support needed at all.
When equipment doesn’t work like expected there is no single minute of good support from “big guys”. Support engineers don’t know anything that isn’t in datasheet (which you as a customer ought to read).

Where is mine support from National semiconductors (is it “big” enough firm), where is from Xilinx, where is from Texas instruments. All that I got from “big” firm was garbage. I was doing state of the art (edge of the technology) not for me of course and all of them only made me idiot… Of course after few months and tons of stupid emails… they all ask me to forgive them and that they are wrong. But what do firm that I work for have from their apologize (nothing, firm that I work for lose lot of money because of the “support”).
Support is myth. You think that you will have support for couple of k$. Only support that lets say Indian engineer read the data sheet instead of you. (and say with Indian accent: “Hello sir, how may I help you”.)

But at the other hand when I made Wishbone to PLB and vice versa bridge and integrate it into Xilinx EDK to be simple like few mouse clicks… They are ask me at the New year day that they have important customer that have a problem with my core. I couldn’t believe my eyes, engineer in the big firm (ASIC designer) don’t know concept of partial decoding. I made a decoder just for him for few minutes and change the core just for him.
The core works just fine according the data sheet that I wrote but the core didn’t do address decoding (it is a bridge not a device that has its address range) but not for the “big” firm engineer.

Thanks,
Petar

[BravoV]

Hey Petar, I'm not upset at all, really,  thats why I asked more details since your goal of making this thread is this below quoted text of yours right ?

I am asking if it would be possible all of us that’s want that scope to make an “organization” or some “community” that will collect money in advance and make a chip oscilloscope that cost at least 4 times less then when you buy it from Tektronix or other “big” firms.

I find its amazing that you're expecting people will gather money for you to start your firm in building the scope with all this posts of yours. Now, instead of lengthy rambling and story telling about your life and your political view which no one interested, why not you do the exact thing that you said you're going to do (read the words in bold below)

Believe me or not some parts that all us buying from Digikey can cost twelve (or higher) times less when ordered in large quantity direct from suppliers and that’s exactly the most priced parts. Of course I can make couple of scope before that and somehow presents to community…

Are you sure you're not from Nigeria ?

[petar]

Thank you, BravoV.

I think people are smart enough to conclude causality.

I am actually doing that, but this is not one day job, give me a time, I am not asking you money right now and expecting you to wait couple of years for it.

Actually, I see that right now I am loosing my time with you and there is better to concentrate on my work.

Petar

>>Are you sure you're not from Nigeria ?


Thank you again.

[EEVblog]

Petar, IMO the only thing that will help sell your scope is analog bandwidth.
That it what traditionally costs money in a scope, and continues to do so.
If you can genuinely do a high performance 500MHz bandwidth scope for a low price, then people will be interested.
Other than that, there would have to some gimmick to rise above the pack, as there are so many USB based scopes on the market, and now companies like Rigol are trying to advance the low end scope market with their new DS2000 series.
And 500MHz probes aren't cheap, they could cost the same as your board...

Also, the devil is in the usability details. If you use a VGA/HDMI output and expect people to use a monitor to use the scope, how do you expect them to interact with it?
Using a mouse seems obvious, but then that's an extra monitor and mouse people have to have on their bench. Many might rightly ask the question, why bother? when I have my notebook/tablet available? In which case why not make it a traditional USB scope?
Do you know how much work is involved in developing a good user interface for a scope?
What about all the fancy stuff people expect these days like serial decoding, masking, fft etc?
That stuff gets easier if you chose a PC as the interface platform, and you could possibly open source the software so people can improve upon it. That's much harder to do if you are doing everything in an FPGA with direct VGA out.

IMO there is room to innovate here in form factor, like for example making it an oscilloscope "backpack" for a tablet perhaps? With real knobs and buttons on the sides perhaps?

Just some food for thought.

Dave.

[petar]

Hi Dave,

Of course the analog BW is what counts…
I could do lets say many GHz BW and let say 20GS/s AD conversion, but for the first scope I don’t want to do that because I can’t do protection circuit (nobody on the planet can at least with current technology) when using inputs 50 ohm to be save to for my users because I expect them to plug input directly to mains which is here in Europe 220Vef.
I am certainly shore that people don’t believe me that I can do 500MHz and what would happen if I claim let say 6 GHz?
If I make few GHz oscilloscope some day it will cost much, not because of BOM but because of user abuse (If something has low cost that users would behave ugly and I don’t want to repair it each time user put input direct to mains…). It needs to cost much that’s how user will think twice before do something foolish.

My previous project is a working oscilloscope. I am not a software expert but I know C and LabView and lot of stuff of windows kernel.

Look at the user interface that I did almost 9 years ago just for me to evaluate performance of my card (I haven’t spent a minute to make it looks good for a customer…).

I already have USB 2.0 IP software and hardware ready for USB – PC based oscilloscope but I don’t want do this way.

One reason is because this way the actual performance depends on PC (I know guy that has 3 windows installed at the same time and 30 programs opened at the same time and never shut down it – always using sleep mode and complain that something don’t work. When use (abuse) it this way it has no chance to work fine and of course he need “support” all the time…).

I could do it (user interface) so easy in Labview or CVI or Measurement studio but this way I have to pay to National instruments for using their software…

I prefer to do traditional benchtop oscilloscope but there is no cheap let say 10” or 12” TFT display on the market (to integrate it with the rest) and it is weird but 20” display is so cheap.

Be shore that I know what I talking about, please believe me that I am qualified.
I know that sound impossible for one man to know and do all the stuff… I know.

Maybe I made a short video on how my previous card work it is a blazing, believe me it a good stuff. I have a camera with just 30 seconds record 640*480….

Thank you for your comments, don’t think I don’t appreciate it if I didn’t comment some of thoughts.

PS. I was thinking of adding voice recognition to be user interface. It could be boring because one have to “teach” the scope commands before using it… From a hardware point of view it is an easy task… From a DSP not so easy, time partitioning, triggering… cross colerration and searching, filtering… Just my thoughts… Mouse is a still an easy solution.

Petar

PS. At the photo (I am learning from you, so photo not a picture, right?) it could be seen flyback DC/DC wave signal on channel1 and output generated by the card itself (DAC) on the channel2.