HP E2310A LogicDart probes - How to build your own probes

[Fraser]

This message was originally posted by me on the HP Agilent Yahoo Group but I thought the information may also be of interest to fellow members of this forum.

Hi to all HP LogiCDart owners or potential owners.

I purchased two LogicDarts recently from the USA. They came without the infamous
probes so were effectively expensive paperweights.

Rumour had it that the LogicDart cannot be used without the HP OEM probes sets
as the probes contained active electronics and all manner of horrors. These
rumours were, to my knowledge, unproven.

Logic Dart probe sets are selling for around $150 per pair so there was an
incentive for me to attempt to build some clones for my own use. I soon
discovered that the LogicDart probe information available on the NET was limited
and mostly incorrect guesses  . People believed HP had placed the high speed
comparators in the probes...this is incorrect. The three comparators are
MAX902's and are located in the LogicDart itself, behind some attenuation and
protection circuitry.
I have been experimenting with a DIY probe solution for a week and have decided to share my discoveries.

This is an 'early days' project but the results to date have been excellent and
I am about to move from the first prototype to a more refined probe for use in
my lab. I welcome any and all comment and if anyone manages to improve upon my
design, please share your data with us so that we may all benefit :-)

Probe connector on LogicDart viewed end on with `flat' at the top

Top Left = Ground / 0V
Bottom Left = Vin
Bottom Right = Vout(active for Resistance,Continuity and diode modes)
Top Right = Not Used

The male 4 pin connector for a probe lead may be constructed using a 0.1" pitch
Molex twin row PCB header or male data connector salvaged from an IDE hard drive
or similar. Such 0.1" pin headers are common inside computers. A shell will need
to be constructed but this is purely a personal choice and hardware task. 10mm
OS diameter Plastic tube with a flat filed on it or a salvaged 3.5mm audio plug
shell could be used.

                          20pF Trimcap
                    |---------| |---------|
Vin -----------|-/\/\/\/\---/\/\/\/\---|-/\/\/\/\------|-----Tip---
                         750k0      36k0           2k0           |
Vout-----------------------------------------------|

0v ----------------------------------------------------- 0v clip

All resistors are 1% tolerance 0.6W metal film types.



LogicDart connector Vin measured dc resistance = 281k5

Measured LogicDart probe series dc resistance = 788k0 (measured)

20pF trimcap required for probe compensation

Total input impedance of the OEM probe + LogicDart is stated as approx 1M / 13pF

The probe series resistance has been measure and tested to confirm that 788k0 is
the required value for correct calibration.  It forms half of a potential
divider, the other half being inside the LogicDart.

Vout provides the drive voltage for resistance, continuity and diode ranges and
was measured at 4V52 dc. I believe this to be a high impedance supply.

The Vout supply inside the LogicDart incorporates a series RF inductor and
GM54GI protection diode (forward biased) in series with the output pin.

The LogicDart probes are passive, containing only resistance and capacitance.
The probe cable from the head to the connector is twisted pair to reduce
capacitance effects.

I do not own or have access to the OEM probes but resistance measurements were
provided by a helpful person who does own one. The readings were confirmed by
practical test using a DIY prototype probe assembly.

Since building this prototype, I have discovered an excellent source of HP probe
information at the URL shown below:

http://www-d0.fnal.gov/trigger/stt/commiss/agilent/agilentprobing.pdf

The resistor values in the document do not map across to the LogicDart probes as
the Logic probes detailed are 100k input impedance rather than 1M. The theory of
construction mirrors my prototype however  

I hope this information is of use and look forward to any comments or
improvement suggestions that members may have. I have yet to test the probe to
33MHz but that will come later when I have a more refined probe build to keep
stray capacitance under control  

Needless to say, I presently have the prototype working very well, and
accurately with one of my LogicDarts. The original outlay for the new Logic Dart
was $105 which compares favourably to HP's original $800 price tag, and for less
than $5 of components it can be made to work without the OEM probes.
Happy days  

[alm]

Good that you got it working (at low frequencies?). If they are passive, you actually may be able to fashion some usable replacements. 33MHz is not that fast, so parasitics shouldn't be too critical. What was the original spec for input capacitance?

The diagram looks screwed up, even in a fixed font. Looks like something like Word tried to be smart and replaced some of the dots with ellipsis. Try converting it to plain ASCII and wrapping it in code tags.

[Fraser]

Hi ALM,

Thanks for your comments.

I had a devil of a job to quickly produce a simple schematic for posting on the HP_Agilent Forum and it all got messed up with left hand pagination and deletion of spaces   I tried to be clever(?) before posting here and added full stops to prevent space deletion issues. I would rather have attached a sketch to my Post and will do so if required for clarity.

You have commented in the past regarding probes and you know the topic well. It was with your previous comments in mind that I considered the bandwidth issues that I might meet when building the final version of the probe. The original spec for the probe tip when connected to a LogicDart is 1MOhm impedance with 13pF capacitance.... that looks pretty poor as a high frequency probe as 13pF presents a pretty low impedance at such frequencies and is likely to cause loading of the source. I was surprised HP accepted such a spec on what was quite an expensive little gadget.

As I have stated, I have yet to fire up my 50MHz function generator to see how the probe behaves but I am expecting some minor issues that will require tweaking    

One issue for me is the cable to use. I understand that most commonly available coaxes such as RG174 present too much capacitance in any reasonable length and scope probes use a specialised thin coax that uses foam dielectric and a very fine centre conductor. I am also aware that some scope probes and logic cables are actually resistive. These issues will be difficult for me to overcome in a DIY build using commonly available COTS  parts. I believe some logic cables are twisted pair to lower capacitance yet be resistant to pick-up and RFI and so I shall try that first.

This is new territory for me but fortunately this is for home use at frequencies below the 33MHz original spec so I am in with a chance of producing useable probes without breaking the bank  


P.S. - I have tried to clean up my original posts wrotten schematic a little   I never was any good at ASCII schematics !

[alm]

Code tags (which force a fixed-width font) usually make ASCII diagrams much easier (edit them in something like Notepad, since the edit form is in a proportional font). Is this correct?

Code: [Select]

                     20pF Trimcap
               +---------| |-----------+
Vin -----------+-/\/\/\/\---/\/\/\/\---+-/\/\/\/\------+-----Tip---
                  750k0       36k0          2k0        |
Vout---------------------------------------------------+

0v ------------------------------------------------------- 0v clip

The original spec for the probe tip when connected to a LogicDart is 1MOhm impedance with 13pF capacitance.... that looks pretty poor as a high frequency probe as 13pF presents a pretty low impedance at such frequencies and is likely to cause loading of the source. I was surprised HP accepted such a spec on what was quite an expensive little gadget.

Yes, the 13pF represents a fairly low impedance compared to the 1Mohm resistor, but is not that unusual. Compare it to 10x scope probes, they're usually 8-15pF or so, even the 500MHz ones (which are not very good at 500MHz unless your source impedance is really low, for this reason). Bandwidth tests are usually done with a terminated 50ohm source, so the capacitor in parallel with the termination resistor just forms a voltage divider with the 50ohm output impedance.

The 13pF is probably just necessary for a properly compensated voltage divider. The ratio between the capacitances should be the inverse of the ratio between the resistors, so if the LogicDart + cable are 30pF (arbitrary number), you need ~11pF in parallel with the attenuation resistor to compensate it. Plus the probe tip will also have some parasitic capacitance (probably even more for DIY construction).

As I have stated, I have yet to fire up my 50MHz function generator to see how the probe behaves but I am expecting some minor issues that will require tweaking   

Keep in mind that your function generator may be less sensitive for both resistive and capacitive loading than a DUT (eg. crystal circuit).

I am also aware that some scope probes and logic cables are actually resistive.

I think all of the Hi-Z probes (for 1Mohm input impedance) (as opposed to the Z₀ probes that are designed for 50ohm) have a resistive center conductor, except for the really low bandwidth or really crappy ones. At least all the good ones designed in the past sixty years or so.

These issues will be difficult for me to overcome in a DIY build using commonly available COTS  parts. I believe some logic cables are twisted pair to lower capacitance yet be resistant to pick-up and RFI and so I shall try that first.

Twisted pair is common for differential signals (since it makes sure that most of the noise that it picks up is common mode, which will be rejected by a differential input circuit), and many pods use differential signaling between the pod and the LA. Were the cables you mention for active or passive pods? That makes a big difference, since active pods have active drivers and can use transmission lines. I don't think twisting does anything to reduce capacitance, but it does limit the loop area between the wires (less susceptible to inductive pickup).

I would probably either use a fairly short piece of RG-174 and add extra capacitance (in parallel with the 20pF trimmer) to be able to compensate it and hope the performance degradation is acceptable, or get the wire from one of those cheap no-name scope probes that you can get for a few dollars. Note that soldering the center conductor is fairly tricky, since it's very fine and doesn't take solder well. One trick I've heard is to use asperin as flux, but keep in mind that melting it produces nasty chemicals (both fluids, which can soak into the isolation, and gases, which you shouldn't breathe in).

[Fraser]

ALM,

Thank you very much for the excellent and helpful information.  Your ASCII schematic is exactly what I wanted to show and is spot on.

The reference to twisted pairs being used in Logic probe pods came from page 9 of the Agilent probing guide already referenced:

"The Standard Probing System
The standard probing system consists of IC clips, probe leads,
probe housing and probe cable. Because it is passive, the standard
probing system is smaller, lighter, and much easier to use
than active probing systems. This passive probing system is similar
to a probing system used on a high frequency oscilloscope. It
consists of an isolation network at the probe tip and a shielded resistive transmission
line. Each probe lead is a 12-inch, twisted-pair cable connected to
the probe cable at the probe housing . The probe tip includes a signal lead, a connector
for a ground lead, and the isolation network."

I now note that it is clearly stated that a resistive transmission line is used
so I will need to see what effect using standard copper will have.

I have read that there are very good reasons why Logic Analyser pods and cables are expensive. In a multi channel situation where matching and propogation times are very important I can see the need. I am just hopeful that the little LogicDart will be a little more forgiving of my efforts  

[alm]

It consists of an isolation network at the probe tip and a shielded resistive transmission
line. Each probe lead is a 12-inch, twisted-pair cable connected to
the probe cable at the probe housing . The probe tip includes a signal lead, a connector
for a ground lead, and the isolation network."

My guess is that this is to reduce inductive pickup, since it's a single-ended input. It's not entirely clear to me if the 'shielded resistive transmission line' refers to the same cable as the '12-inch, twisted-pair cable'. I would expect coaxial construction for single-ended signals.

I now note that it is clearly stated that a resistive transmission line is used
so I will need to see what effect using standard copper will have.

Probably more capacitive loading (not because of the lack of resistance, but because of the larger diameter center conductor in 50ohm cable) and more internal reflections (which are damped by the resistance). Keeping the cables short should help to reduce these effects (you probably don't need 1m leads for something portable like a LogicDart).

I have read that there are very good reasons why Logic Analyser pods and cables are expensive. In a multi channel situation where matching and propogation times are very important I can see the need. I am just hopeful that the little LogicDart will be a little more forgiving of my efforts  

You may be right, because of the limited frequency range.

[Wartex]

6 kits left, 2x probes in each

http://cgi.ebay.ca/HP-E2310A-E2310-61612-LogicDart-Probes-Accessories-/270703133111?pt=LH_DefaultDomain_0&hash=item3f072a01b7

[Fraser]

Hi Wartex,

Thanks for that.

I am already in discussion with that gentleman regarding the purchase of some OEM probes 

Thanks again for thinking of me.

[Wartex]

There was a brand new logicdart auction that ended yesterday, I bid $66 but lost, but with $170 probes, I'm better off buying the Saleae USB logic analyzer. Is there anything about Logicdart that USB analyzers can't do?

[Fraser]

Hi Wartex,

To set the scene for you....

I have been limiting my exploration of all things 'digital' to a decent multimeter, single channel logic probe and an oscilloscope. I am most at home in the analogue and analogue RF domains but times change and I do some basic digital diagnostics as well.


The LogicDart is a well made piece of equipment that is Circa 1998 in design and was considered a clever piece of handheld diagnostic equipment if you could afford it ! It's initial cost was around $800 and at least one reviewer considered it too late into the marketplace and not particularly good value for the money being asked. I have to agree.

I saw the OEM refurbished LogicDarts for sale in the USA for a more reasonable $50 to $110 which is pretty good considering what you actually get in the product and the build quality. When I received my two, they were unblemished on the outside and inside. They have absolutely no signs of use on them and they work well. The big issue with these units is that they use dedicated probes as you will see in this thread. The units come without the probe leads. I believe I have pretty much overcome this issue now and can make my own.

What can it do ? ... there is quite a bit of detail on the Internet that you can read but basically it has the following functions:

Voltmeter, Resistance meter, continuity checker, diode tester, logic state indicator, and most importantly .... a 3 channel logic probe with an 'oscilloscope like' display of the logic states on the three channels, with zoom and capture inspection capability. Timing information and such is available via on screen markers. The logic probe capture function operates at 100 Ms/s which is stated as 33MHz bandwidth capability at the probes. Most logic technologies are catered for such as CMOS, LVCMOS,TTL, ECL etc and custom levels can be set if required. The unit does not have an analogue oscilloscope capability.

Compared to many modern sophisticated multi channel USB logic pods it may seem quite limited and tame. I am using it as a replacement for my trusty old handheld logic probe though, and my needs are quite humble. For me, portability is a plus so the fact that I do not need to use a PC to analyse a data stream is great.

This is a classic example of an item that did not fit the marketplace when launched and failed to sell in significant numbers as a result.  it isn't worth $800! Agilent cancelled production after one year so these units were probably sitting in storage awaiting release as scrap.

[n0gsz]

Thank you for the Logic Dart probe information.  I measured some probe leads, part no. 8120-6909, dated 4797 and 9814 (weird that what are presumably the week and year appear to have been switched around somewhere in the 19 weeks between those dates).  There appears to be a diode in series with the Vout line, with the cathode toward the probe tip and the anode toward the Logic Dart connector.  The diode check function of the Logic Dart shows 0.71 volts of diode forward voltage in that direction and an open circuit in the opposite direction.  The Logic Dart appears to compensate for the forward voltage of one such diode (i.e., the diode in the probe lead being used for the measurement), as shorting the probe used for the measurement to its ground without the probe lead under test in the path shows a voltage of 0.00 volts.  Perhaps you recalibrated the diode check function voltage measurement to your DIY probe lead without a diode, but the Logic Dart I'm using, which came with a complete set of accessories, is calibrated for the probe lead to include the diode.

[Fraser]

@n0gsz,

You are absolutely correct. This thread is now old and I neglected to update it.

I was suspicious that a diode was in the 'Vout' line to the probe but was uncertain whether there was more complex circuitry involved. I did, in fact place a 1n4148 in that line and all continued to work well.

I then managed to buy a several sets of the new probes from a chap in the USA. I tested them and even took X-Ray images of their contents. This was my first access to the OEM probes rather than relying upon others to take measurements on their probes for me. I quickly established that there was indeed a semiconductor junction in the V-out line of the probe assembly. I was surprised to find what appeared to be capacitors in the plug assembly but could not positively identify them.  Since I have three Logic Darts and around 6 complete sets of probes, I stopped developing the DIY probe solution. The OEM probes and the associated 'browser' tools are certainly nice to use. I also bought several of the small IR linked HP thermal printers that often appear on e*ay. They are used with some gas testing probes so appear to be available in decent numbers and at low cost. I paid around GBP20 for my units and they were new in the boxes.

If anyone needs me to do more work on the DIY probe I will consider doing so but there appeared little interest generated by my initial Post beyond the very helpful comments on better design of such a probe.

[n0gsz]

Aside from the scanned printed manual for the Logic Dart downloadable from the Agilent site, this thread is the ONLY information on the Logic Dart I've found anywhere on the internet, so thank you for your work on the probes.  I think I'll make a few simple probes at least for low-frequency applications, as I won't be concerned about physically damaging homemade probes.

I suppose what appeared to be capacitors could be something like MOVs for overvoltage protection or something like that.

I'll have to keep my eye out for an IR printer.  I don't know if the 82240A and 82240B differ in any respect relevant to the Logic Dart.  I see mention of different characters sets and escape sequences, so, unless the 82240A has been confirmed to work properly with the Logic Dart, I may hold out for a 82240B.  Meanwhile, I see someone has written a 82240B printer simulator for which some kind of IR hardware interface can apparently be built to receive IR signals from HP calculators and, presumably, from the Logic Dart:

http://hp.giesselink.com/hp82240b.htm

I see someone else has developed a serial to HP IR interface, of which the optoelectronic portion may be useful for adapting to use as an IR interface for the above 82240B printer simulator.

http://www.wire2wire.org/HP82240B_adapter/HP802240B_adapter.html

While it would be tempting to develop a more comprehensive way of transferring measurement information, including waveforms, from the Logic Dart to a PC, at some point the merit of simply using a USB-based logic analyzer wins out.

[Fraser]

Yes LogicDart information is pretty scarce. I managed to obtain a video of HP presenting the Logic dart to a room full of techies. The presenter went into the detail of how the design came about and the challenges that had to be overcome. It was the first handheld test instrument that HP produced that engaged the talents of its test equipment design team in co-operation with their Calculator design team. That produced the required technical performance whilst achieving an ergonomic handheld device. The presenter also comments on the difficulty of producing the probes and how humidity tests had revealed a vulnerability that they had to address. The Logic Dart was tested rigorously before release but sadly, no matter how neat and well built, it failed due to late arrival into the market and its high price. Trivia...the name LogicDart came for the term objet d'art HP were very proud of their new baby and the success of the two HP teams collaboration on the project.

I may produce some teardown pictures of a Logic Dart if there is interest in such. Its a lovely design but has some weird wrinkled in it.....such as using a flexi cable for connection to the keypad that is a pressure connection with no physical securing what so ever. the un-insulated curved ribbon just presses on the gold contact strip  That had to come from the Calculator design team !

Some people thought that the probes had high speed comparators in them. A theory I quickly proved incorrect when examining the LogiDart. The three comparators are on the main PCB adjacent to the I/O connectors.

I will post some pictures of the units internals (if I can still find them).

Aurora