Transistor Jfet HP equivalent

[HUGOMAN]

Hi to All,

it's long time i look for equivalent of Jfet 1855-0743, even though i found a list by google but it lacks this reference, please indicate with what we could substitue this transistor by near one or what.

Thank you in advance for the support

Best Regards

[dacman]

What is it used in and what is its reference designation?

[rs20]

TLDR: OP wants TO-92 depletion mode JFETs suitable for replacing the current shunt selection JFETs in a 3458A.

Googling for 1855-0743 finds https://xdevs.com/doc/HP_Agilent_Keysight/3458A/doc/3458A%20CLIP.pdf

Which lists description as "TRANSISTOR J-FET N-CHAN D-MODE TO-92" and matching manufacturer part numbers as NSID5092/F50188A . But then the trail goes cold for me, at least as far as finding matching datasheets is concerned.

However, looking at the schematic, these transistors are almost connects to nets with guard rings, which implies that extremely low leakage current would be a requirement. Given that this is a multimeter, high-frequency performance seems unimportant. So here's a shortlist:

http://www.digikey.com.au/products/en/discrete-semiconductor-products/transistors-jfets/288

However, the parametric search here doesn't help with finding JFETs with exceptionally low-leakage at cutoff. Does anyone else know of parts specially designed for this purpose? Also, I can't be bothered tracing the nets through to see what acceptable threshold voltages would be.

Assembly 03458-66501
Q201-Q206: p48
Q107-Q108: p49
Q12-Q13, Q15, Q18, Q24: p50

Assembly 03458-66502:
2 others, didn't look into further.

[dacman]

If you have some that are working, you could measure their conductance and cutoff voltage and then look for a cross.  About all JFETs are D-MODE (depletion mode).

[HUGOMAN]

Thanks rs20 congrats for this details, you reach the target because i am stacked in this position like i have leaky ones, so almost this info i have but i am on a problem with this dmm right now i can't solve the problems that i encoutred on .

Thank you dacman, welcome this idea i thought about but i lack this method, i am make inverse i see other hp dmm using the stages which are similar, but the best if to draw a schematic even by hand to show how can find out the rds and it's conduction of the J-fet.

Best Regards

[djnz]

Rochester Electronics lists 5092_F50188E as an SKU for which they have acquired manufacturing rights.

Perhaps they could have a few samples that / could supply in low quantity.

https://www.rocelec.com/parts/details/?part=587e460ae4b0744c9e039919

[45Overload]

Interesting that the official description is "depletion mode JFET" because all JFET's operate in depletion mode.  I think stressing over the HP part designation is a red herring.  Generic part numbers become mystery HP parts due to hand selection and not specialized production.  That would apply especially to JFET's which vary all over the place from part to part.  I disagree that the guard rings mandate a low leakage JFET. (high Z at cutoff). They are there to prevent interaction between the sections due to pc board conduction and proximity effects.  More likely they were chosen for a reliably low Z at 0 bias.  The JFET's are only used on the part of the string that are relatively high resistance (note use of relay for the low value resistances), so it makes sense that small variations between the JFET switches would not be a big deal. 

[Kleinstein]

In the current shunt switching part the voltages are low and thus leakage would be at a low voltage of maybe 0.2 V or so. This way even quite normal parts could have leakage well below pA.

It still may makes sense to have selected / binned fets, because of the large scattering. So even if the part number for the normal part would known, one might have to check what properties are really needed. Not every fet for that type might fit. One could well end up with a more normal 2N4392 or J112 selected for not too much leakage.

Si JFets are depletion mode. There are a few SiC / GaN JFets that are actually enrichment mode or in between.

[dacman]

Conductance and V-cutoff isn't that hard to check (out of circuit).  If I were tasked with testing conductance and V-cutoff, this is what I might do (if I didn't have a JFET tester):

For Conductance or On Resistance:
1) Clip the negative lead of a DMM onto both the Source and Gate.
2) Clip the positive lead onto the Drain.
3) Measure Resistance.
4) The reciprocal would be the conductance.


For Vg-Cutoff:
1: Get two voltage sources, a 1 kOhm resistor, a 10 kOhm resistor, and two DMMs.
2: Connect one Voltage Source High output to the 10 kOhm resistor High terminal, the 10 kOhm resistor Low terminal to JFET-D, and Voltage Source Low to JFET-S.
3: Connect the other Voltage Source Low to the 1 Kohm resistor High, the 1 kOhm resistor Low to the JFET-G, and Voltage Source High to JFET-S.
4: Set the Voltage Source that is connected to the 10 kOhm resistor to 10 Vdc.
5: Set the Voltage Source that is connected to the 1 kOhm resistor to 0 Vdc.
6: Adjust the 10 Vdc Source Voltage for a 1 mA indication on the DMM that is connected in series with the Drain circuit.
7: Adjust the Vg Source Voltage (which is connected for a negative output) for a 0.001 mA indication or less.
8: The source that is connected to the Gate circuit is now set to Vg-cutoff.

Note: I have selected JFETs in this fashion.

[HUGOMAN]

Thanks Dacman and Kleinstein

Awesome technic, with usefull details, also i Appreciate strongly the communicated of equivalent parts by Kleinstein.

i don't have yet Dca pro, just only i have chiness product it displays for the 1855-0743  N-Jfet  I= 2,0 mA vgs= 1,4v.
i will look for 2N4392 to test it to see what difference between, without forgetting to perform your method to read the rds and vgs.

Regards

[bfsy]

Can this be replaced by 2N4392 ？？？

[Ground_Loop]

If you have some that are working, you could measure their conductance and cutoff voltage and then look for a cross.  About all JFETs are D-MODE (depletion mode).

I did exactly this trying to match some JFETs in my HP 3456. Had to pull a good one and run it on my curve tracer.

[Kleinstein]

i don't have yet Dca pro, just only i have chiness product it displays for the 1855-0743  N-Jfet  I= 2,0 mA vgs= 1,4v.

That sounds close to a J113.  They are not too hard to find.

The combination of Idss of 2 mA and 1.4 V threshold would be more a smaller JFET like J201/J202. The 2N4393 / J113 are larger JFETs with a higher Idss ( ~ 12 mA) when the threshold is at 1.4 V.

In the 3458 ohms circuit it looks like the JFETs don't need to be very low resistance and there also seems to be plenty of gate votlage to turn them off reliably. So a somewhat larger resistance (and thus ideally lower leakage) jFET would be more reasonable, possibly even the rather high resistance 2N4118.

[Wallace Gasiewicz]

As I remember the jFETS in an HP 3456 are used for the most part to turn things on and off. I think the same goes for the 3458.
The important thing is to have the "on" value the same. To me this means the "on" resistance of the FETs should be low and the same on all the FETs. As I recall the resistance from source to drain was about 1.2 ohms, tell me if I am wrong (please do tell me so we do not confuse the issue for the OP). You can get a bunch of cheap 310 jFETS and sort through them. (not from China)
If the FETs are not matched you will get unusual behavior including unstable readings in the least significant digits.

They are switched "off" hard by large voltages, so I do not think small differences in gate response is very important.

[Kleinstein]

1.2 ohms on resistance would be unusual low.  The usual swtiching types are 2N4393 and similar (PN4393,J113, BSR56,...)  with an on resistance of some 30-50 ohms. The higher threchold types like J112 offer slightly lower resistance, but have less headroom to turn off.

Lower resistance types are used only when really low resistance is needed - even the J108 is at some 8 ohms, so 1.2 ohms would be an exceptional large one. Lower resistance also comes with more leakage.

The J310 type would not work the specs are 25 V max., so too low in voltage.

Matching of the on resistance is usually not that critcal, a few 100 ohms more or less would hardly matter in most uses. The R_on of the 2N4117 in the 5 K range can be a problem is some cases.

Today the problem is the rather limited choice of JFETs in a TO92 case. To18 case often works, but often has the case tied to gate and they are a bit more expensive.

[Wallace Gasiewicz]

Thank you for correcting me on the 310 on resistance. I must have been thinking of something else.
They vary quite a bit in their on resistance even in the same batch. Maybe 40 to 180 ohms.

They did work in my repair of a 3456.

[aronake]

Hi to All,

it's long time i look for equivalent of Jfet 1855-0743, even though i found a list by google but it lacks this reference, please indicate with what we could substitue this transistor by near one or what.

Thank you in advance for the support

Best Regards

Did you find some replacement that worked? Or how did you solve (or not solve) this?

[Bill158]

First, before I start this subject again I want to state that I am not interested in debating why I chose this device to replace the HP 1855-0743.  But I will state why I used a 2N5458 as a replacement.
I had a failed device in my 3458A A1 board.  It was in the current measuring circuit and after troubleshooting I found that A1Q204 had very high IGSS.  I had to pull A1Q201 thru A1Q205 to find which FET was bad.  While I had them out I measured the Rds @ 1ma, the Id(off) @ -7v gate voltage ( the logic "off" voltage at the Gate ), the Vgs @ 1ua, and the IGSS @ -7v.  The Rds was about 180-200 ohms.  The IGSS was under what my 576 CT could measure using the "emitter leakage".  The Id(off) was under what I could measure also.  The Vgs was about -4v.  This was for the other 4 1855-0743 devices.  My choice of the 2N5458 was based on:
1.  I had a few on hand along with other N channel J-FET types.
2.  Measuring the 2N5458 it had similar characteristics.  Mainly the Vgs(off) @ -4v.  The Rds was 220 ohms.  I again couldn't measure the IGSS or the Id(off) using the "emitter leakage" position of the 576.
So I reasoned "WTF, let's give this a try and see what happens".  I could have tried other FETs if this didn't work.
As far as I am concerned the 2N5458 works just great.  I did a CAL 10.000E3 and no issues.  ACAL just the same.  I confirmed the DCI as best as I could comparing the readings with another 3458A.  Resistance measurements are good also.  The specs on this HP device are simply NOT available as all have found out.  HP/A/K will NOT sell you this device and their stock is marked "Protected Stock".  HP/A/K will sell you a replacement 03458-66501 board for US$4900!!!  For a home hobbyist I am just as happy as a duck in water!  For those who are in business, go for the HP/A/K repair at what ever the going rates are now.  Just one man's experience.
Bill

[Wallace Gasiewicz]

Since my last post, I think I have learned something.   
 In the HP3456 the aforementioned accursed FET was used in switching and also as a "clamping diode". As I recall I replaced some of the switching diodes (not the ones used as diodes) with common FETs and that worked.
I then obtained a DAS 45 voltage reference and had it Calibrated.  Using the calibrated values, I found out that the low volt (100 mV) scale on my 3456 units would not work correctly.    My Blissful Ignorance was unceremoniously ended.....   
I Believe-------( I am just a hobbyist, not an Engineer )-------
The problems I had with the 100 mV range were related to the use of the FET as a clamping diode with source and drain tied together.  Eventually I procured some
Thompson 2N4117 FETS and installed some in the clamping positions....they worked, These are extremely low leakage FETS with advertised gate leakage of 120 femto amps,   
I do not know of any diodes that are of such a low leakage value. Perhaps there are some but HP used the FET.
They test in my BSide as two diodes as do some of the original FETs. This is just an observation, I do not know it's significance.
I also had to replace the "pre charge" FET.     
At Present I have been able to calibrate the two repaired 3456 units with my General Resistance DAS 45, so I conclude that they are working.   
I have a third board that has the same trouble with low volt range that I have not fixed.     
At Present, my working theory is that the choice of FET for "switching"  is not as critical as the Clamping Function that they are also used for.

There is another extremely low leakage diode (not a FET) used in the protection of the resistance function that I had to replace on one board,  Cycle Collector on For Sale forum has some. As I Recall, this diode caused high resistance values to be off.  The test for this is easy and is elucidated in the manual.

[Kleinstein]

I some cases the same type of JFET us used as a clamping diode (usually towards to positive side) to get some compensation of the gate leagate current. The switches when off get a negative gate voltage and thus the positive side for compensation. So there can be cases where parts from the same series / batch would be preferred.

Changing an input switch may effect the precharge adjustment.

There are low leakage diodes, especially BAV199  (there is also a glass case THT type, that however needs protection from light). The test specs are not very tight, but the typical (nearly all samples)  leakage is really good. The testing for sub pA currents is expensive and not found with lower cost parts.

[schmitt trigger]

Conductance

For Conductance or On Resistance:
1) Clip the negative lead of a DMM onto both the Source and Gate.
2) Clip the positive lead onto the Drain.
3) Measure Resistance.
4) The reciprocal would be the conductance.

Is conductance somewhat related to Idss? The test setup appears similar.

[Kleinstein]

There is a strong correlation between the JFET parameters.  There are tables on the relations in text books and application notes on JFETs ( e.g. Fairchild AN6609).   Ron is about equal to  1/ Gfs0 .  gfs0 is about  2 Idss / Vgs_off.    This is at least a reasonable approximation.

Vgs_off anyway scatters quite a bit between batches and with V_gs_off also R_on and Idss change. From the same series (e.g. J111 to J113), part with the higher Vgs_off will still have the low R_on as Idss goes up even faster than Vgs_off (more like the square).

[Wallace Gasiewicz]

First: Thanks to Kleinstein for the post about Gate Leakage. I think I now understand why same FETs were used.
About the "setup" for R On and IDSS:   
Setup is the same but the voltages are far different.  The voltage (Drain to Source) for IDSS are higher than what is used to measure R On.  Increase the voltage enough and you will reach IDSS. At that point the FET is Saturated and resistance will be much higher, You are out of the Ohmic Region of the FET.

[fmashockie]

Just FYI although I don't see the 1855-0743 listed, CuriousMarc's website has an archive of proprietary HP part #s and their electrical specs.  It is not a complete list, but it has definitely come in handy for me.

https://www.curiousmarc.com/doc-archive