Dennis,
The most challenging aspect of the analog design for me was the overvoltage and? ESD protection of the Ohms converter current source using a minimum of CHEAP off-the-shelf parts.? I spent many months on this design which I will go through in detail in my DMM video tutorial.? ?Some of the design is quite clever in my opinion - (could it be anything else - LOL)
With no cost constraints and modern components, I would have chosen much better op-amps, voltage reference, and ohm reference resistor.? ?I think you are asking how one might soup up the 3468A.? This can be done and I will address this in my video series.? There are certain errors that are introduced by the input bias currents and offset voltages of the op-amps used in the Ohms and? DC circuits.? By proper analysis of the circuits, the important parameters can be identified to tell one how to select a given op-amp for the function desired to minimize drift over time and temperature.? You can only carry this so far of course because the input hybrid will eventually become the limiting factor.? ?Now of course if you wanted to get crazy about this, you could render the drift of the fineline resistors in the input hybrid U102 as a moot point by surrounding it in a temperature controlled environment but that is getting foolish since the A/D tops out at 300,000 counts.
The single most important imporvements that could be made is using a 40K? Vishay S102 series resistor for the Ohms reference and an LTZ1000 for the DC reference.
The design philosophy for the 3478A was different than the 3468A.? The 3468A was intented to be used as a portable bench instrument that could be taken out in the field with an Hp41C calculator for data acquition.? The 3478A was intended to be used as a system meter hence the HP-IB interface.? Plus as a system meter there needs to be an in-guard and out-guard to isolate the analog and digital sections when used as a system meter in a rack of other instruments.? This greatly stepped up the power requirements which made the battery backup option a non-starter.? Besides that, the 78 was NEVER intended to have battery power.? It was from the start conceived to be a rack instrument.
To further differentiate and justify the price difference between the 68 and 78,? marketing insisted (I objected and was over ruled) that the 78 have one more lower range on DC and Ohms.? You will notice that the 78 has a 30mV and 30 ohm range that the 68 does not.? ?All the circuitry and hooks are in the 3468A to have those functions but the firmware will not turn them on.??
Unfortunately,? I have no way to give you to access those ranges on the 3468A.? That would require a firmware change and these is no spare memory in the Intel 8048 and external chips to hold the required code.? If there was space, we would have put some "back door" way of doing it for testing etc. but we could not.? You could "cheat" the instrument into providing those ranges but that would require external jumpers between Input hybrid pins and would require you to be inside the instrument to manually move the jumpers around.? Besides the display would not position the decimal point in the correct place if you did that so other than to play and have fun, it has no practical use.
On the lower DCV ranges, the input impedance is typically around 10^12 ohms or more.? There is always stray capacitance in parallel with that so as charge bleeds into the input and charges the stray capacitance, the instrument measures the voltage on the capacitance and that is the drift you see.? ?The leakage currents that charge these strays are from the? POSITIVE sources of voltage which are more dominant around the front end of the instrument.? You are more likely to find drift upward due to positive leakage currents rather than negative leakages.? ?
Remember? ? ? Q =? C V? ? ?so? ? ? ? ? ??i? =? dQ/dt? =? C? dVdt
? ? ? ? ?hence,? voltage drift:? ? ? ? dV/dt = i / C
If leakage current is? ?i = 1pA? and? C = 10 pF? ?then the drift is? ? dV/dt? =? ?1pA/10pF? =? 0.1v / second
Thanks for the kind words.? It is hard for me to believe that I was in HP Labs 41 years ago designing the 3468A.? In many ways it seems like yesterday.? ?I have vivid and fond memories of those days and the people I worked with.? The KEY project was a new approach taken by Hp to the design and manufacture,? Our team worked closely with all the support groups from the IC, PCB, marketing, industrial design, machine shop departments and the production floor.? ?I made many good friends from all the various departments.? This was not a case where the engineers in the lab were isolated and thru specs over the fence to the support groups and waited from them to return a proto back to us.? ?I and the other engineers had our own work benches and we each built and tested our own designs and interfaced our modules into a central prototype we kept in our area.? We had no "offices" with doors.? Our desks and work benches were all together where we could see and talk to each other without getting up.? Of course that lead to a lot of pranks for stress releif and camaraderie.? ? A close working relationship with the support groups helped the success of the project.? For example,? the PC shop worked closely with us in developing an new coating process for the motherboard PCB to eliminate teflon standoffs (many in the 3455A and 3456A) to insure high impedance at critical nodes with feed-thru components.? This was a big breakthrough that helped keep manufacturing costs low.? Had that aspect of the PCB manufacture not been successful, meeting the design goals would have failed.? ?There were many new innovations that required various departments to pull off successfully and not just from us design team engineers.
Hope that helps you with your questions.
George Hnatuik, PE
