Scope : wait for new models ?

[tautech]

So you don't consider that the SDS2000X 2 or 4 ch models having more than 3x the memory depth and more wfm/s  than the GDS-2204E as worth anything ? 

I think about it this way. The heart of the scope is ADC. ADC has certain sample rate. The more the sample rate, the more expensive ADC is, and the price grows exponentially. Faster sampling also requires faster and/or bigger FPGAs to accept all these samples. Siglent designed SDS2000X scope around an ADC with 1Gs/s sample rate. As we know, ADC requires low pass filter. For 1Gs/s sample rate, the required frequency of the fron end filter is somewhere around 300 MHz. What Siglent does for SDS2104X? They install bigger capacitors, so the filter is now 100 MHz instead of 300 MHz. What the customer can do:

1) Live with incorrect capacitors and have 100 MHz scope instead of 300 MHz.
2) Open it up and replace the capacitors with the correct values.
3) Pay $1,000 to Siglent so that they would install correct capacitors.

You make a # of assumptions some of which are wrong.
The front end filter is not constructed as you think and SW controls the max BW.
The ADC used is 2GSs/s and 1GSa/a is the minimum sample rate.

[NorthGuy]

The front end filter is not constructed as you think and SW controls the max BW.

May be. Hardware control is just the assumption I made watching Dave's tear down video. Each channel has exactly two hand-soldered capacitors.

The method doesn't really matter. The point is that the low pass filter is set lower than it should. I don't want to say that Siglent is the only manufacturer who does this. This was just an example. Every manufacturer is doing this to some extent. This is a historical thing. I am simply suggesting that this practice might be coming to the end because crippling the products makes them more costly to produce and less competitive at the same time.

[Someone]

I am simply suggesting that this practice might be coming to the end because crippling the products makes them more costly to produce and less competitive at the same time.

This discussion has been around a few times now, many markets do this for segmentation. In some markets you pay for exactly the same thing with a different name or sales channel and the prices are wildly different, so be happy test equipment is only priced to its capabilities. There is no evidence that a test and measurement manufacturer would make more money by selling a single model instead of a range of models, or one of them would be doing it and taking the market to themselves.

[tautech]

The front end filter is not constructed as you think and SW controls the max BW.

The method doesn't really matter. The point is that the low pass filter is set lower than it should. I don't want to say that Siglent is the only manufacturer who does this. This was just an example. Every manufacturer is doing this to some extent. This is a historical thing. I am simply suggesting that this practice might be coming to the end because crippling the products makes them more costly to produce and less competitive at the same time.

Yep, it's been going on for decades. While some (but not all) offer after-sales purchasable BW upgrades, but this amounts to an acknowledgement that BW is crippled in SW and fosters/seeds resultant interest in hacking equipment. Manufacturers (most) don't want this added attention and prefer the status quo where a series of equipment is developed for a set cost with the price for the top of the series helping subsidise the lower units in order to get the total sales required for profitability.

[Fungus]

I am simply suggesting that this practice might be coming to the end because crippling the products makes them more costly to produce and less competitive at the same time.

 

100% ass-backwards. They do it because it's cheaper. Cheaper to only have one PCB for all models, one production line making all models, the ability to change the 100MHz:300MHz output ratio at a moments notice (based on incoming orders). All this far outweighs the $10 they'd save by using a cheaper ADC in the 100Mhz model.

Plus: If you're smart you can pretend to let customers 'hack' their scopes.

How many DS1054Zs would Rigol sell if it wasn't soft-unlockable? None - everybody would buy the slightly cheaper GW-Instek 1054B.

How many HMO1002s could R&S sell if it were easily hackable to 300MHz? Zillions! Why don't they allow it?   

[JanJansen]

How many HMO1002s could R&S sell if it were easily hackable to 300MHz? Zillions! Why don't they allow it?   

The HMO has everything unlocked exept bandwidth i think, thats why its nice you dont have to hack it, at least i dont see why i need the bandwidth.

Look i found this site : https://www.soselectronic.com/products/rohde-schwarz/hmo-1002-165549
Anyone know this site ?, its weird, they have only 1 out of many scopes in stock,
als with the DMM they have only 1 UNI-T in stock, weird site, only good price on the HMO1002, why cant i get it cheaper from a "normal" shop ?

[JanJansen]

Its software, you can hack any software, only some hacker has to take the time and money to do this.
If you wanto solder in your new scope, go ahead, i dont like it, you must be some super professional person to say that blueskull.

[pascal_sweden]

Even if you could upgrade the HMO for free, it is still no good.

It's built to a price point. Avoid it! Don't waste your money!

[NorthGuy]

100% ass-backwards. They do it because it's cheaper. Cheaper to only have one PCB for all models, one production line making all models, the ability to change the 100MHz:300MHz output ratio at a moments notice (based on incoming orders). All this far outweighs the $10 they'd save by using a cheaper ADC in the 100Mhz model.

It is not about production, it is about sales.

Company X makes 500 MHz scopes and sell them for $7,000. They also make the same scopes and modify them to be 100 MHz and sell them for $2,000. Apparently, they make profit on their $2,000 scopes, which means that their margin on $7,000 scopes is huge. Such margins may be appropriate for Keysight, but such margins are not going to persist in a competitive market.

Imagine, company Y makes substantially the same scopes but marks everything as 500 MHz and sell them for $2,500. Low end customers would prefer that because they get 500 MHz for really small premium over 100 MHz scopes made by company X. High end customers will get huge savings compared to 500 MHz scopes made by company X. So, company X will have nobody to sell to.

[JanJansen]

I like to start building my own scope, start with the most simple scope & logic analyzer if that is possible, later look for good/fast parts, and find out how they work.

[JanJansen]

I,m only building DIP, so i bet those fast scope parts are not even available in DIP,
i,m curious what is the best scope you can build with DIP packages.

[rstofer]

I like to start building my own scope, start with the most simple scope & logic analyzer if that is possible, later look for good/fast parts, and find out how they work.

It's pretty easy to get wiggly lines on a screen, I did it about 60 years ago using plans from the ARRL Handbook.  I wouldn't want to guarantee the volts/div or time/div but I get neat Lissajous patterns.  The project looked like a DuMont scope as shown in this Wiki:

https://en.wikipedia.org/wiki/Allen_B._DuMont

Of course, mine didn't have anything like the legends on a real scope.  Still, not bad for a youngster.

Today, I would just use a logic analyzer to capture a free running ADC and deal with the details later.  It's pretty easy to build a 200 MHz logic analyzer but that is not an impressive sample rate.

https://www.sump.org/projects/analyzer/

Going real fast with the ADC and FPGA is going to take a bunch of engineering.  Memory design ought to be a challenge!

DIP packages?  You aren't going anywhere.  There are no FPGAs that I am aware of in that package and very few fast uCs.

[nctnico]

The hardware of a scope is pretty easy compared to the user interface software. One of my ongoing projects is the FPGA+firmware part of a network distributed 'oscilloscope' used for data acquisition purposes.

I build my (edit:) first oscilloscope from a LED VU meter kit which also had a linear mode. I added a 16x8 LED matrix which showed the input signal over time by sampling the output of the chip on the VU meter kit. Not long after that I got a real scope for my 14th birthday.

[ebastler]

The HMO has everything unlocked exept bandwidth i think, thats why its nice you dont have to hack it, at least i dont see why i need the bandwidth.

Is that so? Then why does that shop you link to list all the wonderful decoding upgrades for the HMO, priced at €200 to €500 each?
 

Look i found this site : https://www.soselectronic.com/products/rohde-schwarz/hmo-1002-165549
Anyone know this site ?, its weird, they have only 1 out of many scopes in stock,

Indeed, an unusual selection of stock items. They seem to be a subsidiary of Conrad Electronic, mainly for Eastern Europe, hence are probably for real. If you are in the market for a Hameg/R&S scope, it can't hurt to ask them for their prices on the non-stock models?

[JanJansen]

The hardware of a scope is pretty easy compared to the user interface software.

Programming is the fun part, i used to making games.