New Agilent scopes

[Leo Bodnar]

But as you said, you can probably roll your own, just putting 90k resistors arrays on your test boards and using a 80pin ATA cable that has extra ground wires should work fine.

Thinking again 80 way ATA cable is not necessary because original 40 pin ATA did not have interleaving GND between signals that's what 80 way was introduced for.  Plugging standard 40 pin ribbon cable into Agilent probe socket will have GND interleaving automatically.

So there is no auto-sensing for the digital pods plugged in? Unless one of the side GND pins is used as an active low signal.

[Rufus]

I lashed up a cable and 'pod' today. About 42cm of ribbon from an old IDE cable and 2cm of stripboard.

I measured the impedance looking into the ribbon and MSO. It is 10k at dc, down to about 2k5 at 1MHz, about 250R at 10MHz, dropping to about 60R at 70MHz then bounces up and down with standing waves above that.

I guessed at 220R followed by 91k with 6p8 in parallel as an attenuation/compensation network. The impedance of that was about 17k at 1MHz, 2k at 10MHz, and drops smoothly towards about 250R above that.

I put 4 x 5 way 0.1" headers on the 'pod' because I have loads and they are cheap for making custom adapters. 4 signals and a ground on each. The scope end stiffened with a bit of fibreglass and hot melt glue. I attach some photos and what the scope made of a 300mV 20MHz sinewave on the 'pod' (digital threshold set to 0).

[EEVblog]

There are tiny wires woven into the fabric. Amazing!

The resistors are in the end of each probe cable, not the pod.

Yes. If they were in the pod then the bandwidth would be very poor due to the cable capacitance.
Just like how x10 scope probes work.

Dave.

[EEVblog]

Thanks for the photos. Now I don't have to open mine.
I always just assumed they had some buffer circuitry in the pod, never bothered to measure it.

Dave.

[tnt]

@Rufus: Very nice !

How did you measure the impedance ?
You don't happen to have an active probe to show the signal after the compensation network do you ?
I guess the misalignement of the two wave form is a delay caused by it ?

[Leo Bodnar]

I lashed up a cable and 'pod' today. About 42cm of ribbon from an old IDE cable and 2cm of stripboard.
...
I put 4 x 5 way 0.1" headers on the 'pod' because I have loads and they are cheap for making custom adapters. 4 signals and a ground on each. The scope end stiffened with a bit of fibreglass and hot melt glue. I attach some photos and what the scope made of a 300mV 20MHz sinewave on the 'pod' (digital threshold set to 0).

Looks good!  It might make sense to remove unused unterminated wires on the cable edges.

Why do you have capacitors on the probe side?  What inline resistors have you used?

Cheers
Leo

[Leo Bodnar]

I guess the misalignement of the two wave form is a delay caused by it ?

It might be just due to input comparator hysteresis.  Seems to be 150mV from the pictures.  Mind you the sine signal is only 0.3Vpp

I love these scopes.  I have done a good development session yesterday.  I have used CAN bus, SPI and USB triggering all on the same day.

By the way I am not sure if you figured it out but you can get all the 14 days trial licences at once, install them and keep rolling the scope time back to keep them active.

[mikeselectricstuff]

I guess the misalignement of the two wave form is a delay caused by it ?

It might be just due to input comparator hysteresis.  Seems to be 150mV from the pictures.  Mind you the sine signal is only 0.3Vpp

I love these scopes.  I have done a good development session yesterday.  I have used CAN bus, SPI and USB triggering all on the same day.

By the way I am not sure if you figured it out but you can get all the 14 days trial licences at once, install them and keep rolling the scope time back to keep them active.

I'm guessing they probably self-destruct if they see a date after expiry, so could be annoying if you forget...

[bandtank]

I guess the misalignement of the two wave form is a delay caused by it ?

It might be just due to input comparator hysteresis.  Seems to be 150mV from the pictures.  Mind you the sine signal is only 0.3Vpp

I love these scopes.  I have done a good development session yesterday.  I have used CAN bus, SPI and USB triggering all on the same day.

By the way I am not sure if you figured it out but you can get all the 14 days trial licences at once, install them and keep rolling the scope time back to keep them active.

How do you get them all at once?

Couldn't you roll the time back several years and never have to deal with it or will that not work?

[Rufus]

@tnt
I measured impedance with a cheapy vector network analyzer. I didn't look at the signals after the network.

The scope probe was about 90cm longer than the cable+pod so you would expect maybe 8ns offset just from that. Hysteresis on the digital comparators would look like delay. The digital signal not being square would be voltage offset somewhere. It would be easy enough to accurately measure the digital input offset and hysteresis with a low frequency triangle wave but I haven't done it.

@leo
The values were in the post, 220R then 91k with 6.8pF across it. Look at the discussion of probe impedance in the user manual, the real thing has something like that in it. The apparent impedance of the ribbon cable drops rapidly with frequency so high frequency attenuation would be huge without some capacitance bypassing the main series resistance. You pay the price with lower probe input impedance. The 220R puts a limit on it.

[Leo Bodnar]

How do you get them all at once?
I'm guessing they probably self-destruct if they see a date after expiry, so could be annoying if you forget...

You just request them on the same day so that they will all be active during the same period.
If they lapse you will need to go back in time and simply install them again from the flash drive.

Couldn't you roll the time back several years and never have to deal with it or will that not work?

It will complain about this and ask you nicely to advance the time.

[Leo Bodnar]

@leo
The values were in the post, 220R then 91k with 6.8pF across it. Look at the discussion of probe impedance in the user manual, the real thing has something like that in it. The apparent impedance of the ribbon cable drops rapidly with frequency so high frequency attenuation would be huge without some capacitance bypassing the main series resistance. You pay the price with lower probe input impedance. The 220R puts a limit on it.

Agilent actually makes generic breakout probes 40pin -> 20pin and 40pin to 40 pin woven probe cable.  Well done Agilent!

You have pretty much nailed it exactly right.
By the way the probe schematic is on the page 6 of http://www.teknetelectronics.com/DataSheet/HP_Agilent/HP_01650_63203.pdf
Or page 9 of http://www-d0.fnal.gov/trigger/stt/commiss/agilent/agilentprobing.pdf.  This is an excellent review of probing solutions and techniques.

[Rufus]

By the way the probe schematic is on the page 6 of

Thanks for the links I hadn't seen those. Optimum compensation must depend on the cable and I expect standard ribbon is lower impedance than the Agilent weave. If I were guessing again I would probably guess at 10pF.

[Leo Bodnar]

Thanks for the links I hadn't seen those. Optimum compensation must depend on the cable and I expect standard ribbon is lower impedance than the Agilent weave. If I were guessing again I would probably guess at 10pF.

Agilent says the cable impedance should be about 80-100 Ohms preferably towards 100 Ohms.  Standard IDC cable is often quoted as 100 Ohm impedance so it's not such a bad solution.
Weaved cable is very flexible so it maybe the main reason for Agilent using it?   There is also "high flex" version of standard IDC cable but it's quite expensive - $15-$20 per metre.  Still cheap compared to Agilent prices.

[Hypernova]

Weaved cable is very flexible so it maybe the main reason for Agilent using it?   There is also "high flex" version of standard IDC cable but it's quite expensive - $15-$20 per metre.  Still cheap compared to Agilent prices.

Well your standard IDC cable is unlikely to survive all the twisting from daily use. Look at how much effort goes into those twisty phone cables.

[mikeselectricstuff]

I have a vague recollection of reading a while ago that the woven cables used nichrome wire - just measured mine & it measures 200R, which would be stupidly thin if it was copper....

[bandtank]

Weaved cable is very flexible so it maybe the main reason for Agilent using it?   There is also "high flex" version of standard IDC cable but it's quite expensive - $15-$20 per metre.  Still cheap compared to Agilent prices.

IDC cable is not suitable for logic analyzer probes in any capacity other than basic hobby work. Woven cable is the only choice if you need to do real work. It's expensive for a reason.

[Leo Bodnar]

IDC cable is not suitable for logic analyzer probes in any capacity other than basic hobby work. Woven cable is the only choice if you need to do real work. It's expensive for a reason.

Wow, you should consider going into politics!   
Seriously though, I have a woven Agilent set but I'd like to have a special kit for direct plugging into boards at development stage.

[bandtank]

IDC cable is not suitable for logic analyzer probes in any capacity other than basic hobby work. Woven cable is the only choice if you need to do real work. It's expensive for a reason.

Wow, you should consider going into politics!   
Seriously though, I have a woven Agilent set but I'd like to have a special kit for direct plugging into boards at development stage.

lol. I work on systems with up to 5000 channels and many people have tried to buy cheap cables or come up with clever solutions to avoid the cost of woven cables. It really never works for a variety of reasons and we always end up buying the good ones. I understand what you are saying and it is nice to have cables that plug in directly. I use woven cables for that as well though. Obviously it depends on the complexity of your system which is why I said it's fine for basic hobby work. I don't want to wonder if my system is broken or if my instrument is possibly to blame as my designs are typically 5-8GHz, so not using good equipment is a waste of time and effort.

[Leo Bodnar]

Here is an eBay listing that has a full monty for $229.  You get two full sets of probes, two breakout boards plus some box with a screen you can make aquarium from.
http://cgi.ebay.com/HP-1651B-Logic-Analyer-w-Accessories-2924A00437-/320233764679?pt=BI_Analyzers&hash=item4a8f6b7747#ht_1537wt_907
I'd had it but I am in the UK.

Why nichrome wire?  It's pretty rigid.

[oPossum]

More HP porn....

The unmarked resistor is 90.0k

[oPossum]

Thanks for the photos. Now I don't have to open mine.
I always just assumed they had some buffer circuitry in the pod, never bothered to measure it.

Dave.

I also assumed they where active probes. It was a surprise to find nothing at all in the pods.

The amazing attention to detail of HP/Agilent products is one of the reasons I buy old HP gear rather than the latest cheap black box with a USB port. Look at the probes on a cheap USB logic analyzer and compare to almost 20 year old HP tech.

I have an old HP 54645D (the original MegaZoom), and the only thing that I would consider replacing it with would be a new 3000 series. That won't happen until I can get a used 3000 at a price I can afford. Until then the ability to turn a knob and zoom in 10,000 times continues to amaze and delight.

[Rufus]

Why nichrome wire?  It's pretty rigid.

For the resistance to make a lossy transmission line. If you search you will find Agilent have a patent for it.

[tnt]

@Leo: Ah well, I ended buying another kit (exactly the same model as oPossum posted) for 250USD incl shipping. (I'm in europe as well so shipping would have killed me on a full logic analyzer anyway :p)

@oPossum: Wow you sacrificed one of the probe. Nice

* tnt can't wait to get his 3034 tomorrow :p

[mikeselectricstuff]

More HP porn....

The unmarked resistor is 90.0k

Presumably the little track stub on the left is a spark gap for ESD protection - I've seen similar on TV external inputs.