Why no medium priced oscilloscopes ?

[theoldwizard1]

Admittedly I have been "out of the loop" for a long time !  The first 10-15 years of my career I used a 'scope almost daily.  I have not touched one in probably 30 years !

As we all know there are some really good low priced scopes on the market these days.  There seems to be a huge jump in price from these units to the "professional" model (Tek, etc.)

There are some features that also seem to be missing:

• True differential input (or maybe I should say isolated ground ?)
• >8 bit resolution A/D (10 or maybe 12 seems MUCH better)
• >500 MS/sec when all channels are in use (reasonable glitch capture requires about 10 sample per repeating time period)

I would think that all of those feature could be added for a $100-$200 dollar.  Of course the question always is, will the public pay that ?

Comments ?

[exe]

• True differential input (or maybe I should say isolated ground ?)

This one I don't know how much effort to implement. But would love to have this a standard option. I'd want isolated inputs...

• >8 bit resolution A/D (10 or maybe 12 seems MUCH better)

I propose to skip 10 and 12bit devices and go for 24bits as it could serve as good 8.5DMM as well . I think the cost of a good frontend and fast 10+bit ADC is prohibitive, that's why only few vendors provide true 10+bits of resolution, and most of them use ASIC.

• >500 MS/sec when all channels are in use (reasonable glitch capture requires about 10 sample per repeating time period)

Most modern scopes are based on jelly-bean 4-channel 1Gs/s ADCs. That's why they are cheap. For two active channels it will provide 500MS/sec. Anything better than that would cost much more due to two factors: economy of scale and cost of development. Most companies figured out how to make 8bit 1GS/s 100MHz-200MHz scopes, there is a lot of competition and prices are very affordable (comparing to what was, say, 5-10 years ago, afaik). Going much beyond that requires a lot of effort and raises cost of BOM sharply.

Nonetheless, I hope one day we'll have devices matching performance of, say, rtb2004, but more affordable.

UP: what do you call "affordable"? For me it's around $600, for others it's $2000, for others it's below $100... It also seems that people who _really_ need high-performance scopes do have budget for them. So, while we may demand 1+Ghz scopes for the price of dso1054z, do we really need them or it's just TEA?

[joeqsmith]

I am not sure what medium priced means to you.  For me, medium would be $3,000 to $15,000.   Adding a couple hundred wouldn't be a factor. 

I would rather have the single ended inputs than differential.    Say I am wanting to look at 3GHz differential 5V signal....  Then I wan't to look at some diver stage in a 400V SMPS.  That's not going to be easy...   Consider too that the cost of a differential probe could very well be as much at a medium scope (using my numbers).    There are just so many variables.

I use a scope because I want to see what things look like.  If I need higher resolution than what the scope offers, I am normally using a bench meter or some other sort of dedicated acquisition system.

500MS/s, no problem.

[theoldwizard1]

Most modern scopes are based on jelly-bean 4-channel 1Gs/s ADCs. That's why they are cheap. For two active channels it will provide 500MS/sec. Anything better than that would cost much more due to two factors: economy of scale and cost of development.

My expertise WAS in embedded microprocessor systems so I know next to nothing about analog design.

But using one of the major distributors online tools, I stumbled across the Analog Devices AD7933.  4 channels, 10 or 12 bits, 1.5 MSPS for LESS THAN $10 !

OP: what do you call "affordable"? For me it's around $600, for others it's $2000, ...

Good point.  For sure, 12 bits, 4 channels @ 500 MSPS, true differential input, under $1000.   2 channels, 10 bits @ 500 MSPS, no differential input, under $600.

[theoldwizard1]

I am not sure what medium priced means to you.  For me, medium would be $3,000 to $15,000.   Adding a couple hundred wouldn't be a factor. 

That is WAY beyond what I would consider "medium" priced, especially for a non-professional.

I don't know what the capabilities of a $10,000+ digital oscilloscope are, but if you are looking at 3GHz signals, I think you will need an analog 'scope !

I use a scope because I want to see what things look like.  If I need higher resolution than what the scope offers, I am normally using a bench meter or some other sort of dedicated acquisition system.

Obviously !

Like I said, my background was in embedded microcontroller system, mostly automotive.  We had no "front end" except to maybe clamp high voltages.  Early on we choose a 10 bit ADC.  Of course it was thrifted down to 8 bit.  Then the reference voltage was dropped from 10V to 5V.  Not the best situation as the "ambient noise" took up a larger proportion of the total bandwidth.

[janoc]

I think some of the newer Siglent scopes could do what you want, with the exception of the differential/isolated inputs.

I would call that a fairly niche feature that would add a lot of costs for no reason if all you want is to poke  around an SMPS where a cheap 20MH high voltage differential probe will do the job just fine. If this was built-in, it would have to be on all channels and e.g. 200MHz capable instead. Massive overkill few people would use or pay for in this low end price category.

Having a 50 Ohm termination built-in as opposed to the usual 1MOhm is probably going to be more useful for most people than having differential or isolated inputs.

[joeqsmith]

I am not sure what medium priced means to you.  For me, medium would be $3,000 to $15,000.   Adding a couple hundred wouldn't be a factor. 

That is WAY beyond what I would consider "medium" priced, especially for a non-professional.

I don't know what the capabilities of a $10,000+ digital oscilloscope are, but if you are looking at 3GHz signals, I think you will need an analog 'scope !

I use a scope because I want to see what things look like.  If I need higher resolution than what the scope offers, I am normally using a bench meter or some other sort of dedicated acquisition system.

Obviously !

Like I said, my background was in embedded microcontroller system, mostly automotive.  We had no "front end" except to maybe clamp high voltages.  Early on we choose a 10 bit ADC.  Of course it was thrifted down to 8 bit.  Then the reference voltage was dropped from 10V to 5V.  Not the best situation as the "ambient noise" took up a larger proportion of the total bandwidth.

I am still not sure what medium priced would mean to you.   High end scopes to me is normally $15K to $100K.   It's why it is better to provide details rather than use ambiguous terms.   

I gave away my only analog scope several years ago.   I had bought it brand new back in the 80s.  HITACHI, something 100MHz 4 channel, with readout.  Used a Z80.    My first digital scope was an old HP logic analyzer that had two scope channels. 

You can see three of my old DSOs looking at a 3+ GHz oscillator I cobbled together a while back.  The first picture I show is my first real digital scope.  It's really old, released in 1989 so 30 years old now.   I would hate to think how much of our tax dollars it took to buy one back then.     The other two are still pretty old.       
https://www.eevblog.com/forum/testgear/nice-older-dso-advice/msg2450235/#msg2450235

As they age, the prices drop and they get in the range of the home hobbyist.    Just depends what you are after. 

[JxR]

I am not sure what medium priced means to you.  For me, medium would be $3,000 to $15,000.   Adding a couple hundred wouldn't be a factor. 

That is WAY beyond what I would consider "medium" priced, especially for a non-professional.

Honestly, for a mid-tier scope I would say that price range is very conservative.  Personally, I would say the mid-tier range is closer to $7k to $30K.  Is that expensive, hell yes it is!  But, mid-tier pricing and capabilities has nothing to do with  our own personal budgets.  My scope on sale was around $2.5k (MSRP ~$6k), and it is considered low-end/entry level.  Is it a very good entry level scope? Yes, one of the best in my opinion (Rhode & Schwartz RTB2004).

"Medium price" is a subjective term, and it is based on an unknown price range that has personal meaning to you, but no one here really knows what that is unless you tell us. You can help the experts (I'm not one of them) here out, by defining your budget. 

There are scopes with differential inputs, PicoTech makes one specifically for the automotive industry, it starts at ~1.2k. 
https://www.picotech.com/oscilloscope/4444/picoscope-4444-overview?lv

There are also portable battery powered scopes like those made by MicSig (so no earth reference on battery).  These seem popular for automotive applications, and start around ~$600.
http://www.micsig.com/html/86.html 

You can also get active differential probes for any scope, that has already been mentioned.  Expect to pay anywhere from ~$150-$2k (or more) depending on what you need.

There are also tons of options for older analog scopes on the used market by Tek and others.  You can definitely get a great analog scope easily under ~$1k.

[bdunham7]

Why do you want more bits? 

The availability of products like these depends on both demand and the cost of making them.  I don't think most users at the low end level (or even most professional and high-end uses) would benefit from more bits.  And then you need a front-end and ADC with noise levels low enough to justify more bits--which is not likely for an entry-level DSO.  And more processing power, memory, etc etc. 

As to differential inputs, that is a niche feature that is actually quite costly to implement.  Even the lowest-cost Chinese differential probes are well over $100 for one channel.  If you think a full-bandwidth 4-channel fully isolated differential front end could be implemented at some low cost, go ahead and make one--you'll be able to sell the interface alone for more than a nice scope.  There are already nice handheld units with 4 isolated inputs, although they cost a few thousand dollars.....

...which brings me to the thing you said that surprised me the most.  Why, having the experience you describe, do you consider $600 to be a "medium priced" oscilloscope???  What did a "medium price oscilloscope" cost 30 years ago? Entry level models like the DS1054Z absolutely demolish units of similar function from 30 years ago that cost 10 times the price, not even adjusting for inflation.  The new MSO5000 series would seem to me to be the very bottom end of "medium priced" (just under $1K and up) and will be a fantastic bargain if they ever get the wrinkles ironed out. 

[tggzzz]

Arguably the question can apply to medum priced anything.

In my first job, a long time ago a sales representative made the point that he knew how to sell things for £100 and £10000, but not for £1000.

£100 was under the corporate radar and sold like consumables from petty cash; he didn't invest personal time in selling it, they just sold.

£10000 was a capital purchase requiring management approval, and he could invest time personally demonstrating the widget.

£1000 fell between the two stools: to expensive for petty cash, too cheap for him to demo it.

[ResistorRob]

I get what you're saying. I've been researching what oscilloscope to buy for over a year now, so I have to completely agree their is a huge hole in middle range area. I commented the same thing over on the Tek 4 Series thread. For $500 you get a really solid scope and sometimes spending $1,000+ will actually buy you less features because you often get a base model that needs more than the original cost for software upgrades to unlock things like serial decoding.

I was really excited when the Rigol MSO5000 ($1k) was released but they purposely put in a dim screen that nobody is going to want as to not eat sales away from their 7000 line.

I LOVE the Rhode & Schwartz RTB2000 and when it was released they had a promo deal to get all the features unlocked for free. Sadly I didn't have the money at the time, so later on when I had some money I looked into and the price would be at least $2,000 more without the launch special pricing. I heard the rumor about an all in one pricing to get all features unlocked coming out soon. I'm out of work right now, so definitely won't end up on my bench, but I commend R&S for listening to feedback from potential customers.

If Rigol or Siglent or anyone really wants to dominate the mid-range scope market (which I define in the $1k-2k range) all they need to do is take their $400-500 scope and add a touchscreen that is responsive with a good UI (and obviously not dim!), 50 ohm, and all the little things that hamper beginner scopes price it at $1,700 with everything included and they would sell like hotcakes.

I know on this forum not many people consider $2k mid-range but when you can buy five $400 scopes, or four $500 scopes for the same price that is a significant difference. Imagine applying that same pricing concept to cars... You wouldn't say a car that costs more than 450% as a Toyota Camry still in the same category! I really want that color palette feature from the new Tek scopes in something under $2k... I would grind hard to come up with the cash to buy something like that. A boy can dream, can't he :-)

[exe]

I stumbled across the Analog Devices AD7933.  4 channels, 10 or 12 bits, 1.5 MSPS for LESS THAN $10 !

Wait a second, it's not the device you want in your oscilloscope. Remember you are asking 500MSPS per channel, I fail to see how a 1.5MSPS devices can be useful in this case. If you need a digitizer, DMM6500, it's a 6.5 digit dmm with a 16bit 1msps digitizer.

If 1.5MSPS is enough for you, check Analog Discovery 2 (AD2), it has, among many other features, two differential inputs  (14bit, 100MSPS ~30MHz, afaik, check datasheet). I have one and I really love it. A swiss army knife device, one of a kind (okay, may be not, there is ADALM2000, but I don't think it has a good software support as AD2, which crucial).

PS most scopes have a "high resolution mode" which also helps sometimes to get 10bits or even more... Works by oversampling, you trade speed for resolution. But it needs enough noise to work. Also there is averaging, that helps sometimes too. I'd say an 8bit adc cannot compete with a true 10+bit adc, but may come close.

[tautech]

Admittedly I have been "out of the loop" for a long time !  The first 10-15 years of my career I used a 'scope almost daily.  I have not touched one in probably 30 years !

As we all know there are some really good low priced scopes on the market these days.  There seems to be a huge jump in price from these units to the "professional" model (Tek, etc.)

There are some features that also seem to be missing:

• True differential input (or maybe I should say isolated ground ?)
• >8 bit resolution A/D (10 or maybe 12 seems MUCH better)
• >500 MS/sec when all channels are in use (reasonable glitch capture requires about 10 sample per repeating time period)

I would think that all of those feature could be added for a $100-$200 dollar.  Of course the question always is, will the public pay that ?

Comments ?

If this was easy there would be quite a range of scopes with isolated channels to choose from.

It ain't !
To do it properly every I/O needs full isolation plus there are signal fidelity issues to address too.
Even the probes used on an isolated channel scope are NOT some run of the mill basic scope probe in order to give the operator some decent protection.
https://store.siglentamerica.com/product/pp830-300-mhz-400-v-oscilloscope-probe-used-shs1000-series-models/

For the moment cheap differential probes are the way to go but some are limited for BW.

Siglent have an optional ISFE module that gives full isolation for 2 channels at just $135 but it's only 1 MHz BW rated.
https://store.siglentamerica.com/product/isfe-isolated-front-end/
Don't expect it to fit any scope unless you use BNC cabling as the BNC spacings for the BNC-BNC adapters are different to most other scopes.
Recent discussion about ISFE here:
https://www.eevblog.com/forum/testgear/siglent-sds1204x-e-released-for-domestic-markets-in-china/msg2444670/#msg2444670


On the matter of sampling, two 1 GSa/s ADC's as used in the SDS1104X-E does offer 500 MSa/s with all channels active.

[theoldwizard1]

There are scopes with differential inputs, PicoTech makes one specifically for the automotive industry, it starts at ~1.2k. 
https://www.picotech.com/oscilloscope/4444/picoscope-4444-overview?lv

That 'scope is GROSSLY over priced !  Over $1200 !!  It is based on 15+ year old technology !  Some of the specs:

Bandwidth limit 1 MHz
USB 2.0 (which limits the continuous sample rate to 16 MS/s)
4 channel sample rate of 100 MS/s
256 MS total

[David Hess]

• True differential input (or maybe I should say isolated ground ?)

Differential inputs were a rare built in feature even in the past and even then, differential probes were available.  Today this feature is provided by differential probes.

Using two channels in add and invert mode can work with a modern DSO although usually not as well as an analog oscilloscope because of increased quantization noise.  (1) This is one place where increased digitizer resolution would make a significant improvement but most people just use a differential probe.

Isolated inputs are a different thing and very useful.  Many modern DSOs support this but it is somewhat of a specialized feature.  I suspect most users who require a differential probe would do better with isolated inputs.  Compared to a differential input:

- Built In
- Lower Noise
- Higher and More Reliable Common Mode Rejection
- Higher Bandwidth
- Low Common Mode Impedance - This is worse than with a differential input but not usually a problem.

• >8 bit resolution A/D (10 or maybe 12 seems MUCH better)

Greater than 8 bit resolution is of limited use.  Vertical channel accuracy, linearity, settling time, and noise usually make 8 bits more than enough unless bandwidth is severely restricted.

• >500 MS/sec when all channels are in use (reasonable glitch capture requires about 10 sample per repeating time period)

Lots of DSOs meet this requirement.  Even lots of 30 year old DSOs meet this requirement.

I would think that all of those feature could be added for a $100-$200 dollar.  Of course the question always is, will the public pay that ?

It is not about the cost to implement these features.  It is about what customers will pay for them.

A majority of customers will pay that much for a change in style and features which make a DSO worse.  Low contrast user interface and controls?  They have that.  Want a glossy screen which reflects the room lights into your face?  You are in luck.

(1) Some early DSOs used analog channel switches so do not suffer any increase in quantization noise making them particularly suited to add and invert differential operation.  These DSOs are a little weird; when the common mode rejection is adjusted in add and invert mode, you can visibly see the trace noise eerily decrease.

[theoldwizard1]

There are scopes with differential inputs, PicoTech makes one specifically for the automotive industry, it starts at ~1.2k. 
https://www.picotech.com/oscilloscope/4444/picoscope-4444-overview?lv

Actually, the automotive verision os the Pico 4425.

Similar specs, but at least it has BNC connectors !  It does have some nice software for your PC, but again, GROSSLY overpriced.

[David Hess]

For the moment cheap differential probes are the way to go but some are limited for BW.

High voltage differential probes always have limited bandwidth compared to an isolated input when common mode rejection is taken into account.

Even the probes used on an isolated channel scope are NOT some run of the mill basic scope probe in order to give the operator some decent protection.

Even old isolated probes were good about safely insulating the operator making them look like something built by The LEGO Group.

[guenthert]

Oscilloscopes are good at showing squiggly lines and not much more.  For that 8bit is enough, well with larger and higher resolution screens, more would be nice of course.  Seems that Rohde&Schwarz sees a market opportunity there.

Hmmh, perhaps worth mentioning the Digilent Analog Discovery again.  It got a 14bit ADC (and differential inputs).  For squiggly lines that would be overkill, but the Discovery is a multi-function device and can also act as a spectrum analyzer, for which 8bit certainly is not enough.  At 100MS/s it might not be suitable to find glitches (depending of the duration of those glitches of course) and its triggering performance is only so-so, but for the intended market (educational and hobbyist) I think it's a good value.

Don't expect a medium-prized 1GS/s version of it anytime soon though.  Digilent is now a National Instruments company ...

[JxR]

There are scopes with differential inputs, PicoTech makes one specifically for the automotive industry, it starts at ~1.2k. 
https://www.picotech.com/oscilloscope/4444/picoscope-4444-overview?lv

That 'scope is GROSSLY over priced !  Over $1200 !!  It is based on 15+ year old technology !  Some of the specs:

Bandwidth limit 1 MHz

20Mhz Bandwidth...but who's counting.  If it was "grossly" overpriced, then no one would buy it.  If this type of scope had a large demand, then other manufacturers would move in to sale a similar scope at a market acceptable price.  They've been selling it for years, so I'm assuming there must be a market for it.  Simple economics. 

I personally have no use for such a scope.  Again, why not just give us your budget?  The thing is this, if you can't find what your looking for in your budget range, then it basically means you need to relax your requirements, or increase the amount your willing to spend.  If what you used to use "back in the day" provided you the functionality your looking for, then have you tried finding whatever that was on the used market?

[David Hess]

I thought Yokogawa had a differential input DSO but if they did, based on their web site it is discontinued.  They do have several isolated input instruments.

[bdunham7]

Here's a system that sort of halfway starts to meet the OPs requirements!  2 isolated channels, 500+ MSa/s per channel (in ETS or "repetitive store" mode).  Still 8 bits though. 

All probably $10K worth back when new, still worth perhaps $1K today.  Both units have 4-channel equivalents for more money.

Not what most people would consider useful or a good value proposition for ordinary use, although the isolator comes in handy sometimes. 

[maginnovision]

I got you buddy. https://www.ebay.com/itm/Fluke-192-ScopeMeter-Oscilloscope-Scope-Meter-VERY-NICE-SHAPE-WITH-CASE-ETC/382981392685

It's a monster, it's got some bandwidth, deep 27.5k memory/ch, 500MS/s sampling rate, isolated inputs and an 8-bit ADC.

[theoldwizard1]

Hmmh, perhaps worth mentioning the Digilent Analog Discovery again.  It got a 14bit ADC (and differential inputs).  For squiggly lines that would be overkill, but the Discovery is a multi-function device and can also act as a spectrum analyzer, for which 8bit certainly is not enough.  At 100MS/s it might not be suitable to find glitches (depending of the duration of those glitches of course) and its triggering performance is only so-so, but for the intended market (educational and hobbyist) I think it's a good value.

The Analog Discover 2 is based on the AD9648 and has 1MΩ, ±25V, differential, 14-bit, 100MS/s (capable of 125 MS/s), 30MHz+ bandwidth.  Save a few bucks and the 12 bit part, AD9628.

Put 4 AD9628 in parallel, clocked together, and you have most of an "honest" 125 MS/s 'scope.  Add the front end from the Analog Discovery 2 and your almost home.

[theoldwizard1]

If it was "grossly" overpriced, then no one would buy it.  If this type of scope had a large demand, then other manufacturers would move in to sale a similar scope at a market acceptable price.  They've been selling it for years, so I'm assuming there must be a market for it.  Simple economics. 

The automotive specific version, 4425, has many features for auto techs.  One of the biggest features if you sign up for their service is storing of "know good" waveforms from various sensors on various engines.

Again, why not just give us your budget? 

Under $1000

4 channels
10/12 bits
differential input (or maybe just isolated ground)
100 MHz front end, 500 MS/sec on each channel
1 Mpts storage

This is very close to the Keysight DSOX1204A except for the 10/12 bits and the differential input

[maginnovision]

The scope you want doesn't exist, you're going to have to make some compromises. Unless you just want those to be the specs whether they're real or not.