50 ohm inputs for Siglent SDS1104X-E

[backyard_cnc]

My new Siglent SDS1104X-E lacked the 50ohm inputs that are required for most of the work I do in RF communications so I ordered up a pair of adapters from Amazon when I ordered the scope. They showed up today and a quick test with my tinySA ultra as a 10mhz signal generator shows that the -18dbm signal with its calculated conversion of 79.6mv p-p measures as 80mv which is significantly closer than the measurement without the impedance matching especially when the output of the tinySA is not necessarily precision calibrated!

[David Hess]

The adapters just have a 50 ohm resistor in parallel with the oscilloscope input, which still has its associated input capacitance of about 15 picofarads.  They are useful to 200 or 300 MHz.

[rf-loop]

All we who have old times calibrated and repaired Tektronix analog oscilloscopes are familiar with "Input Normalizer" (and there was many of them due to different inputs.  Yes purpose was different but principle same.   

This principle can use also for external 50 ohm termination (Disadvantage is attenuation..)  as well explained in this tube.. 

[skander36]

On a Micsig scope (TO 3004) I have obtained even a slightly a better rise time with this load resistor(P57) than with it's internal 50 ohm path.
The result was 1,060 ns for internal 50 ohm and 0,993 ns with P57. With another source the result was  1.606 ns external 50 ohm(P57) and 1,879 ns for internal 50 ohm.

[Aldo22]

I also bought one of these, but I think these are quite expensive for what they are.
So I made one myself for the second channel using two SMA connectors and two SMD 100Ω resistors as well as some copper shielding.
I can't measure any difference to the commercial product. Is there one?

[tautech]

LOL
You need a VNA, then you would see !

[Aldo22]

LOL
You need a VNA, then you would see !

Sure! I meant theoretically, something I hadn't considered?
Those Cleqee p57s don't look like rocket science to me. 
I mean, that's two resistors, nothing else substantial.

[xrunner]

My new Siglent SDS1104X-E lacked the 50ohm inputs that are required for most of the work I do in RF communications so I ordered up a pair of adapters from Amazon when I ordered the scope. They showed up today and a quick test with my tinySA ultra as a 10mhz signal generator shows that the -18dbm signal with its calculated conversion of 79.6mv p-p measures as 80mv which is significantly closer than the measurement without the impedance matching especially when the output of the tinySA is not necessarily precision calibrated!

Amazon Review of product P57 -

3.0 out of 5 stars Does not meet advertised specifications

Reviewed in the United States on December 24, 2023

Verified Purchase

When this device is swept from DC-1GHz the return loss is really quite terrible when UHF is approached and the device is actually unusable above 800 MHz, and even that is a generous compromise. I wouldn't use them above about 500 MHz and only at very low voltages. Not worth it.

These might be okay for very low voltage low frequency termination, but that's about it. Check out B07G566JC7 for a budget solution that actually works properly.

I intend to return this product. They are really not good even at this price and do not come close to meeting the advertised specification.

https://www.amazon.com/Zyyini-Adapter-Through-Terminator-Device/dp/B07SD3TC51/ref=sr_1_2?crid=X32P7JBA4PNO&keywords=p57+load&qid=1704025061&sprefix=p57+load%2Caps%2C103&sr=8-2

[tautech]

LOL
You need a VNA, then you would see !

Sure! I meant theoretically, something I hadn't considered?
Those Cleqee p57s don't look like rocket science to me. 
I mean, that's two resistors, nothing else substantial.

Only a balanced transmission line. 

[Roger Need]

I bought several of these P57 adapters to use with my Siglent 1202X-E.

They work fine for me up to 400 MHz with this scope.

Here are some tests I did on two of them with a NanoVNA.  Return loss looks good to me and the traces from the two overlap well so there is not a significant difference between the two units.

Roger

[joeqsmith]

Another thread discussing this topic.
https://www.eevblog.com/forum/testgear/inexpenive-50-ohm-feed-through-terminations/msg5227260/#msg5227260

[the_cake_is_a_lie]

Is there any advantage of using those adapters over a BNC T adapter with one end fitted with a 50 ohm terminator? That's what I have and the combination is slightly cheaper. Maybe the adapters are more convenient and that's enough justification.

[CharlotteSwiss]

I bought several of these P57 adapters to use with my Siglent 1202X-E.

They work fine for me up to 400 MHz with this scope.

Here are some tests I did on two of them with a NanoVNA.  Return loss looks good to me and the traces from the two overlap well so there is not a significant difference between the two units.

Roger

My 1202 is 200Mhz, but I see there are 1202 that go up to 400Mhz? 

[tautech]

Charlotte, while a 200 MHz DSO might display a 400 MHz signal the amplitude will be very reduced so measurements of such will never be representative.

[CharlotteSwiss]

Charlotte, while a 200 MHz DSO might display a 400 MHz signal the amplitude will be very reduced so measurements of such will never be representative.

I thought not even an over frequency signal would detect it. Ok thanks for the information 

[Veteran68]

Is there any advantage of using those adapters over a BNC T adapter with one end fitted with a 50 ohm terminator? That's what I have and the combination is slightly cheaper. Maybe the adapters are more convenient and that's enough justification.

It's not the same, but depending on what you're measuring, it may or may not make a critical difference. If you're not into sophisticated high frequency RF, the T-adapter terminator is probably good enough.

Here's a link to a recent video of Joe's where he's comparing different through terminators and mentions the drawbacks of the T-adapter termination. Relevant content is at 3:55 into the video.

https://youtu.be/gy10caH95Qg?si=o80vL_-NN5GuyMo-&t=235

[Njk]

I bought several of these P57 adapters to use with my Siglent 1202X-E.

I've three. To be known, a maintenance on arrival was needed for all of them, to make them actually usable. The round cup was hard to rotate but easy to get it stuck. A considerable force was necessary to overcome the friction, causing excessive wear to the metal parts. This is not how a good BNC connector supposed to work. A tiny drop of gear oil in proper place solved the problem. Perhaps the boxes of more expensive brands are lubricated at the factory while in this case it's left to users. Not a big deal though

[the_cake_is_a_lie]

Cool thanks. My interpretation is under 50 MHz that I'm interested in, it doesn't matter. At some high enough frequency, the termination method makes a difference and being hardcore and wanting PFC on a probe or scope's parasitic capacitance, maybe you got to DIY. If we say 1 GHz has a 30cm wavelength, a 3cm or so extension on the T splitter is enough to generate reflections and we might wonder how close to 50 ohms the resistor really is.

[joeqsmith]

Cool thanks. My interpretation is under 50 MHz that I'm interested in, it doesn't matter. At some high enough frequency, the termination method makes a difference and being hardcore and wanting PFC on a probe or scope's parasitic capacitance, maybe you got to DIY. If we say 1 GHz has a 30cm wavelength, a 3cm or so extension on the T splitter is enough to generate reflections and we might wonder how close to 50 ohms the resistor really is.

I would say that's basically going to be true. 

Personally for high speed designs, I most likely am using specialized probes and scopes.   The thru terminations mentioned would never come into play. 

While I never came right out and stated it, that video with the SPICE simulations was more just to show why the original Pasternak thru terminator that used axial resistors out performed my original ones made with surface mount parts, once we considered the effects of the scopes loading.   I assumed most picked up on that when I was presenting the data.   Maybe not..

[alm]

Cool thanks. My interpretation is under 50 MHz that I'm interested in, it doesn't matter.

Just be careful that this is 50 MHz bandwidth, not a 50 MHz repetition rate. If you are looking at the rising edge of a fast 50 MHz square wave, that may well require much more bandwidth than 50 MHz and hence be much more sensitive to termination.

[joeqsmith]

Just be careful ...

True.

I suspect Rodger's 30dB @ 700MHz return loss was without the terminators attached to their scope.  This is what graybeard had suggested I measure way back when.   Just took me a few years to get around to it.  The video I made walked through the basic math and simulation to show why this is and how it could potentially be improved. 
 
https://www.eevblog.com/forum/testgear/50-ohm-inputs-for-siglent-sds1104x-e/msg5252313/#msg5252313

[MathWizard]

So SMD resistors don't have much parasitic capacitance, but is it mainly that, which screws up the P57 response past 800MHz or less ?

[David Hess]

On a Micsig scope (TO 3004) I have obtained even a slightly a better rise time with this load resistor(P57) than with it's internal 50 ohm path.
The result was 1,060 ns for internal 50 ohm and 0,993 ns with P57. With another source the result was  1.606 ns external 50 ohm(P57) and 1,879 ns for internal 50 ohm.

I would assume that the difference is caused by peaking because of inductance in the resistor or layout.

Is there any advantage of using those adapters over a BNC T adapter with one end fitted with a 50 ohm terminator? That's what I have and the combination is slightly cheaper. Maybe the adapters are more convenient and that's enough justification.

There will be a difference at higher frequencies, but likely no difference at 50 or 100 MHz.

I bought several of these P57 adapters to use with my Siglent 1202X-E.

I've three.

I have collected two different types over the years.  I always wanted one of the small Tektronix ones as a standard for performance, but ended up with a pair of the larger 5 watt Tektronix ones which are only good to a lower frequency, and one of the 1 GHz ones like that discussed here.

[joeqsmith]

So SMD resistors don't have much parasitic capacitance, but is it mainly that, which screws up the P57 response past 800MHz or less ?

If you followed along in that video where I went over some of the basic math and simulate the termination,  I am suggesting it is the scopes input that screws up the response, which I assume is why graybeard asked my to look at it.   

[the Chris]

If you followed along in that video where I went over some of the basic math and simulate the termination,  I am suggesting it is the scopes input that screws up the response, which I assume is why graybeard asked my to look at it.

I enjoyed watching the video a lot, great presentation of measurements and their evaluation.

My takeaway was that for best performance, a custom termination is adjusted to the individual scope on which the termination is going to be used. I especially liked the use of an inductance to linearize the response. Felt like adusting a speaker crossover.

Thanks for the recording,
Christian

[skander36]

On a Micsig scope (TO 3004) I have obtained even a slightly a better rise time with this load resistor(P57) than with it's internal 50 ohm path.
The result was 1,060 ns for internal 50 ohm and 0,993 ns with P57. With another source the result was  1.606 ns external 50 ohm(P57) and 1,879 ns for internal 50 ohm.

I would assume that the difference is caused by peaking because of inductance in the resistor or layout.

Very likely, because the signal looks quite peakey 

[joeqsmith]

If you followed along in that video where I went over some of the basic math and simulate the termination,  I am suggesting it is the scopes input that screws up the response, which I assume is why graybeard asked my to look at it.

I enjoyed watching the video a lot, great presentation of measurements and their evaluation.

My takeaway was that for best performance, a custom termination is adjusted to the individual scope on which the termination is going to be used. I especially liked the use of an inductance to linearize the response. Felt like adusting a speaker crossover.

Thanks for the recording,
Christian

I was planning to make a terminator that I could trim.   Ran some simulations for it but didn't have parts on-hand to construct it.   OP who started the last thread had lost interest, so I saw no point taking it any further.   

I summed it up in the end.  For best performance, buy tools capable of making the measurements you need to make.   You may be able to gain some performance by hacking up a custom terminator like I show but if you are working with say high speed digital,  it is going to be difficult to match the performance of a better scope and probes.   Of course, that can be six figures which is why you see old vintage equipment in my home lab.   

[David Hess]

I was planning to make a terminator that I could trim.

When Tektronix first started making 1 GHz oscilloscopes, they discovered that all of the 50 ohm BNCs they could buy were not 50 ohms, and ended up making their own.

A trimmed terminator is not going to matter unless the cables, connectors, and adapters are consistent with it, and suitable precision resistors are available.

[joeqsmith]

I was planning to make a terminator that I could trim.

When Tektronix first started making 1 GHz oscilloscopes, they discovered that all of the 50 ohm BNCs they could buy were not 50 ohms, and ended up making their own.

A trimmed terminator is not going to matter unless the cables, connectors, and adapters are consistent with it, and suitable precision resistors are available.

That's the idea of using a trimmer, we are compensating for these small errors.  Similar to why you need to compensate all your 10X probes for the best performance. 

I don't know what these precision resistors would be for that you mention. 

We can get some idea of what Tektronix was up to in the 1950s, looking at the release of the 519.  It appears to use 125 ohms with N125 connectors. 
https://7vmc31.p3cdn1.secureserver.net/wp-content/uploads/2020/07/HighSpeedScopeBrochure_507_517_519_1961.pdf

Looks like you can still find some of these parts at salvage yards.  125ohm thru terminators.   Nice. 
https://www.surplussales.com/EQUIPMENT/TEKHPLISTED.HTML

[David Hess]

I was planning to make a terminator that I could trim.

When Tektronix first started making 1 GHz oscilloscopes, they discovered that all of the 50 ohm BNCs they could buy were not 50 ohms, and ended up making their own.

A trimmed terminator is not going to matter unless the cables, connectors, and adapters are consistent with it, and suitable precision resistors are available.

That's the idea of using a trimmer, we are compensating for these small errors.  Similar to why you need to compensate all your 10X probes for the best performance. 

I don't know what these precision resistors would be for that you mention.

I mean 50, 100, and 200 ohm 0.1% surface mount resistors are available, and 50 and 100 ohm 0.1% RF surface mount resistors.

We can get some idea of what Tektronix was up to in the 1950s, looking at the release of the 519.  It appears to use 125 ohms with N125 connectors. 
https://7vmc31.p3cdn1.secureserver.net/wp-content/uploads/2020/07/HighSpeedScopeBrochure_507_517_519_1961.pdf

Looks like you can still find some of these parts at salvage yards.  125ohm thru terminators.   Nice. 
https://www.surplussales.com/EQUIPMENT/TEKHPLISTED.HTML

The 519 was a special case where the oscilloscope deflection plates were driven directly without a vertical amplifier, and Tektronix even had a variation of the GR-874 connector with a smaller center connector which was 125 ohms for it.

The Tektronix 011-0129-00 was a 0.1% 50 ohm feedthrough termination but only good up to 100 kHz.  As I recall it was for calibration of something.

The Tektronix 017-0084-00 was a 50 ohm GR-874 air line used as an absolute impedance reference of 50 ohms +/- 0.4%.

[joeqsmith]

...
The 519 was a special case where the oscilloscope deflection plates were driven directly without a vertical amplifier, and Tektronix even had a variation of the GR-874 connector with a smaller center connector which was 125 ohms for it.
...

Indeed it was special as it appears to be Tektronix's first GHz BW product.  I take it your first comment wasn't actually about their first GHz products but something much later.

When Tektronix first started making 1 GHz oscilloscopes, ...

No matter.  It all very old tech and not anything relevant today.   Even my 4GHz 80's LeCroy is in that category but it does use a standard 50 ohm input w/ SMA. 

[David Hess]

When Tektronix first started making 1 GHz oscilloscopes, ...

No matter.  It all very old tech and not anything relevant today.   Even my 4GHz 80's LeCroy is in that category but it does use a standard 50 ohm input w/ SMA.

It came up with the 7A29 1 GHz vertical amplifier in the 7104 1 GHz mainframe.  The amplifier was designed with a BNC input and Tektronix found out that commercially available BNCs were not actually 50 ohms, so they made their own which was.

I doubt SMA connectors had this problem, but I wonder if TNC or N connectors did.

It is relevant today if you are expecting BNC connectors to be 50 ohms.

[joeqsmith]

..
It is relevant today if you are expecting BNC connectors to be 50 ohms.
..

I thought the discussion was about low end scopes with no internal 50 ohm termination.  The dominant error will be the scope's front end, not the connector.  Still, for the interconnect, while my 80's LeCroy used SMAs, the one from 2000 uses a their own interface.  No BNCs.   

But I am sure people who work in this are understand that even bent coax will cause a mismatch. 
https://www.eevblog.com/forum/rf-microwave/is-is-possible-to-repair-a-hairline-crack-in-semi-rigid-coax/msg4896009/#msg4896009

***

... I doubt SMA connectors had this problem, but I wonder if TNC or N connectors did. ...

All connectors have their limitations, including SMAs.  I have a few 3.5mm connectors at home.  These are similar to an SMA but are an airline.  The following article talks about the history.  It includes a plot comparing the SWR for combinations for 3.5mm and SMA interconnects. 
 
https://www.gore.com/sites/default/files/2016-06/Intermateability-SMA-Connectors.pdf

I wouldn't be at all surprised to find some of these connectors costing more than the scopes we are talking about in this thread.   

[backyard_cnc]

The main focus of my need for these is 50ohm work under 50mhz and besides with a 200mhz BW on the scope I’m hardly concerned about how this device might perform at such high frequencies!