Simpson 444 Micro-ohmmeter Teardown

[don.r]

As requested...

Simpson 444 Micro-ohmmeter 4.5 digits 20mOhm (1uOhm resolution) to 20Ohm range. 100uV peak to peak with a 5mA maximum test current. Can do a 40Hz pulsed or DC test signal with polarity reversal



4 screws on the feet and the bottom lifts off, handle is easily removed to reveal...



Bottom cover



Maxim LCD driver w A/D converter (date code 33/2003?)



Front end, several LT1001 op amps, LF412ACN input op amp, lots of Dale precision resistors



Dale transformer for processing sense line? Date code 35/2003?



User manual including theory of operation: https://www.simpson260.com/downloads/simpson_444_user_manual.pdf

[Vgkid]

Thanks for the teardown.

[don.r]

Another one on eBay "for parts not working" $99 + shipping...

https://www.ebay.com/itm/Simpson-Micro-Ohmmeter-444-Micro-Ohmmeter-w-probes-manual-UFIX/264575648529?hash=item3d99f00311:g:fIEAAOSwSydd~-fE

[don.r]

I have acquired another 444. It's about 13 years older than the first one but the circuit boards are nearly identical.

Older one

[don.r]

Newer one

[Vgkid]

Did you remove the batteries from both of them?
Looks like nothing needed updating over the years

[don.r]

Did you remove the batteries from both of them?
Looks like nothing needed updating over the years

Yes. Both packs were growing whiskers. They use 4 sub C NiCd in series so I'll replace both sets. Yep, no significant change in over a decade. I guess it just works so no need to change it.

[Vgkid]

I was going to say that it was surprising that it didn't need a lot of power , but remembered it ran a very low current(ac) with a lot of gain.

[don.r]

An interesting tidbit, going by the serial numbers (around 30000 for the older unit and 40000 for the newer), it appears Simpson made about 800 units on average every year for those 13 years. Still sold today with a street price near $2000.

[Vgkid]

Seeing that, I wonder how many milliohm meters are still being sold each year?
I'm curios which model was sold the most, they aren't very common.

[don.r]

Not a lot I would think given that a $1500 DMM like the Keithley DMM6500 can go down to 1 uOhm resolution so there is plenty of competition. I haven't seen a meter that matches the ultra-low current/voltage of this one however.

[Vgkid]

Agreed, that simpson dissipates a really low amount of power in the DUT.
Agreed on the modern bench meters, though.
The most modern milliohm meters I can think of are the Hioki's, and the applents. Both way out of my price range.

[jchw4]

Some time ago I acquired this meter (looking exactly like don.r posted, with missing knob).
It had blown input (I think Q5 was open, I just replaced all 4 Q2-Q5) and unstable output.

I attached scope to the input at R33 and after poking around noticed that pressing IC9 changed the output. So I just re-soldered all the connections and now it's much better.
Got down from 150 digits to 15 digits fluctuating.  But still not great, ICL7129 datasheet specifies that  it should be stable to the last digit. Hmmm...

Measurements seems to be about correct, but I don't have calibrated resistors to check.

I thought that it's noisy power rail (I measure about 10mV RMS ripple in AC mode at L1-L2), but in DC mode it's only ~1-2mV ripple, though output is even worse.
It's still ~200 digits fluctuation in DC mode.

Surprisingly, if I short input (with improvised short), result is fluctuating 5-20 (last digits, others are 0). But if I supply external power to L1 and L2 (and disconnect the battery), fluctuation goes down to 0..10. Surprisingly I don't see any ripple change with my DMM, so probably it's some spikes that I don't know how to check.
I checked capacitors in-place, but I don't see anything obviously wrong. C27 in-place has about .8 ohm ESR, so it seems to be good.

Could somebody share their fluctuation? Is this normal? Did anybody see the behavior of a brand new calibrated meter?

I also have question about missing knob. Are they all missing it? Googling reveals that everybody is looking for it and no one has ever found replacement.

Could somebody with the knob provide photos and measurements?

I also have question about DC mode. Could somebody explain how it is supposed to work? I expected that it's symmetrical to the AC, but it's not.
In DC+ mode, Q3 and Q4 are open, Q2 and Q5 are closed which is half of the AC mode. But in DC- mode, all Q2-Q5 are closed, and Q14 (schematics says it's Q12, but on my board it's Q14 that has 10k to D14 instead of direct connection) connects + and - sources, so input should be zero. I traced it and checked schematics, and it looks like it is by design. Or am I missing something?

(The photo is edited: I wiped some wires to clearly see the layout.)

[jchw4]

I feel that something is wrong with the calibration as with shorted input in all ranges source current is about 70% of the specified in the manual.

I would greatly appreciate if somebody can provide their voltage measurements:
1) 20 Ohm range: TP1 to TP2.
2) 2 Ohm range: TP1 to TP3.
3) 200 mOhm range: TP1 to TP4.
4) 20 mOhm range: TP1 to TP5.

Looking at the resistor values and measurement current from the user manual I would expect that measured value should be exactly 1V in all ranges (with shorted input).
(1V also matches ICL7129 specs.)

But in my case it's only ~700mV  .