New designed HP 3450A/B Ohm converter card

[madao]

Hello people.

https://www.eevblog.com/forum/buysellwanted/wtb-(eu)-ohm-converter-for-hp-3450a-b-multifunction-meter/

In this url, i had looked after this card /  good schematic  (avaible scan is very terrible )
 Ohm_converter_original_schematics.pdf (212.86 kB - downloaded 72 times.)

my notice by reserve engineerging (sorry, some is german)
 sammelnotiz für Widerstandmessoptionen_07.07.2021.pdf (1696.13 kB - downloaded 78 times.)

finished schematic, no fault in  digital side.  analog side needs some improvement. (Improvement is finished, TL071 replaced by OPA132 and similar)
 REV_A Schematic.pdf (132.06 kB - downloaded 57 times.)


It works (logic side ) well, only analog side hat a bit problem with high restiance range ( 1M and 10M),  value is  non linear. (~2% )
 I have removed all 5 Diode on OP-amp. With  Diode is  result much worser. 
Despite nonlinearity on higher range is stabilty well.

EDIT: 11:30 , problem with nonlinarity is solved: TL071 should replaced by  OPA132 or similar with extremely low input current.


I make think about puplished of kicad project.  I have  four PCBs left,  who want it. (donate and shipping cost from EU (germany))
[ Specified attachment is not available ]


After finished of pcb, your HP 3450A/B should be modified ( A4 pcb module,  Ohm config wire should be cutted and place with capacitor )

[Kleinstein]

The circuit is quite complicated and cunfusing with the extra floating supply and reference.

The 1N4148 diodes have quite some leakage. The fast gold doped diodes have quite some leakage. The doides D3,D4,D7,D8 should not cause much problems with leakage, but D6 could. This could be a problem for the high resistance ranges. There are lower leakage ones if needed. A transistor base to collector junction or a JFET like the BF245 would be an option for very low leakage.

The TL071 may not be the best choice for low DC noise. I would consider an OPA202 as a low bias, precision BJT based OP.
Capacitve loading to the OP could still be a problem. So stability against oscillation can be tricky.

The BF245 and 1N4148 diode do no provide very much protection. There are higher voltage diodes (easy) and instead of the JFET there are now high voltage depletion MOSFETs available (e.g. BSS129, CPC3980 and similar). Chances are the original used a high voltage JFET that is now rare.

[madao]

Yes,  TL071 is not best chocie, but i have over 100 pieces TL071.  I have hoped, that TL071 is enough.
I would replace it by other better OP, but i must buy it (OPA132 is my farvoit).
BJT-OPV is no chocie, it need really op amp with low input bias/offset current.  HP didn't make a joke with JFET in their distrect op amp.

HP 3450A uses in restiance measuring nonstable 9V z diode reference  instead internal superstable 10v reference. It is enough, because drift of z-diode is complete compensated by ratio principle. This is why, schematic is confused and ciruit is floating.

D6 (for fast op amp recovery) , you're right. It has major influence. I have thought: I should use  B-E of  bipolar transistor, but it is a bad idea.
But  four other Diode on OP amp has also influence.

orginal "high voltage" JFET  has only 50V breakdown.  Better a bit protection than nothings.

Thank for review and critic.

EDIT: I have replaced TL071 by used OPA2132 with adapter ->  nonlinearity in higher restiance is away !

[Kleinstein]

The OPA202 is a very low bias BJT based OP - not all the way as an LT1012, but also not that much higher. The input bias of the TL071 is not that much lower. So chances are it can work, though the extra bias can give some offset and this may no be included in the ratio way to measure the reference. So a true FET type may indeed be better.

The OPA132 is good, a slightly more modern OPA145 would have slightly better specs. It probably would not make a large difference.

The B-E junction of transistors has often only a maximum votlage of some 5-8 V. The BC junction is higher voltage.

[madao]

I have read  datasheet of OPA202 and i must change my opinion with BJT-op amp.  (sorry german)  Das ist geile Käfer, gar nichtmal so teuer.  Thanks , Kleinstein :-)

Yes,  only 5-8V maximum of E-B of BJT is why, it is a bad idea.

Here a gallery of check with  high restiance range. It looks good. Much better with OPA2132.

[madao]

@Kleinstein.

Thank you for proposal with OPA202.

It is very good, a bit better than OPA2132 (w. Adapter)
I have check it today.