HP 4328A Milliohmmeter

While working with the No. 6 Dry Cell and with Power Pole connectors measurements made with the HP 34401A Multimeter using the HP  11059A - Kelvin Test Leads gave erratic readings for measured values below some tens of milliohms.  The reason being that all the dissimilar metal junctions were small batteries whose voltage was temperature dependent.  So for measuring below say 50 or 100 milliohms an AC test method is the way to go and that's what this meter uses.

The test connection uses 5 terminals.   Four terminals are the normal Kelvin connections and there's an additional connection for the feedback that maintains the drive at a constant current.   The control circuit also causes the meter to peg at full scale when the resistance is too high for the range being measured.

Since the test method is AC based all the test connections in this meter are DC blocked with non polarized (film) capacitors so the input is rated at 150 VDC as safe.  This is great since it allows measuring the internal resistance of batteries while they are active or dead or anywhere between.  The newer version of this meter, the Agilent 4338B is only specified for 42 Volts.  I think this is the number that relates to operator safety rather than instrument capability.  So this older meter is far better for testing batteries for tube equipment.

There are limitations to how much inductance can be in series with the resistance to be measured.  The meter is specified to measure a resistance in series with an inductive reactance that's three time the resistance.

This inductor is 280 uH and the 4328A can not measure it's resistance.  Since there's 13" of 22 AWG wire which can easily be measured when it's not wound on the toroid.  The inductive reactance of 280 uH at 1 kHz is 1.75 Ohms.  The resistance of the wire is about 18 milli Ohms or a ratio of 97!  much too high for the meter to seperate.

When making a Kelvin DC resistance measurement the lead lengths do not matter.  But when making an AC resistance measurement the inductance of the leads becomes important.
XL = 2 * PI * F * L so for a 1 milli ohm inductive reactance at 1 kHz you would have an inductance of 159 nH.    A rule of thumb is 1 nH per mm of length for a wire so 159 mm is 6.25".  But the 4328 includes a phase sensitive detector so is supposed to only measure the real part but my guess is it can not do it if the magnitude of the inductive reactance is three times larger than the resistance.  Note that the 1 milli ohm and the 90 ohm R - L networks shown in Fig 5-1 and 5-2 have the inductive reactance twice the value of the resistance.  The test leads that I've seen for sale on eBay look like they are 3 feet long, which is too much for the 1 milli Ohm range.  Another possible reason for the lead length limits by range is the error caused by the difference in length of the test leads prior to joining the Kelvin connections.

Photo at left is showing the resistance of a 50 Amp shunt that produces 50 mv full scale, i.e. it's a 1 milli Ohm 4 terminal resistor.
The A current and B current leads are connected to the high current lugs and the A voltage and B voltage terminals are connected to the small screws.

Above I said the proof was in the pudding, well this pudding tastes very good.  The meter has not been adjusted or tweaked in any way, it just turned on indicating 1.00 milli ohms.  When switching to the 3 milli Ohm range the needle drops back to 1/3 scale right on 1.00 milli ohms.  Sweet.

Note this is a measurement that the HP 34401 can NOT make because it's a DC Ohms instrument.

Now to decide which Kelvin clips to get.

HP 4328A Milliohmmeter Kelvin Test Base

27 April 2008 - This is a wood base made using a couple of plastic 6-32x1" screws and some brass hardware.  A lot lower in cost, and more importantly available within a hour or so of thinking about it.


I have made up some wires for the 5590BA ver 2 Battery Adapter that have the socket pin on one end and a Power Pole 45 Amp terminal on the other end.  The wire is 14 AWG super flex.  Both ends have been joined by crimping.  Before pressing the pins into the socket (very difficult to get them out) I'd like to test them.


I tried to just put the metal terminals under the thumb nut, but the plastic screw bends when it's tightened and that squirts out the round socket pin.  By putting a second (dummy) lead on the other side of the post the forces are balanced and you can easily clamp both leads.  Since the dummy lead is not connected to anything it does not effect the measurement.

The lead shown is reading  about 1.1 Milli Ohm (just off scale on the lowest range.  Note this includes some resistance for the socket pin and Power Pole terminal, the two crimp joints and about 4 inches of 14 AWG wire.


The test leads are the same ones shown above connected to the 50 Amp shunt.

The wooden base is from Michael's (about 60¢) and the plastic screws from Friedman Brothers (about 60¢) plus a few dollars of gas (about $4.00 per gallon).

Mueller 75K Kelvin Clips

Each Jaw is a seperate contact to the two clips shown represent a four terminal (Kelvin) connection.
If you clip onto the edge of 10 mil brass shim stock as one clip is slid along the edge the resistance changes in a very smooth manner.  Three inches of Phosphor Bronze 10 mills thick and about 400 mills wide is about 3 milli Ohms.

The metal may be silver with gold plating.  The problem with the leads as shows is that the max distance between the two test points is limited to a few inches.  This is a good thing for good accuracy the Amp Turns for the loop formed by the test leads is minimized, but it's to too convienient for some measurements.  Another set of leads with more reach would be better.  Fir example the HP 1159A Kelvin Test Leads can reach a little more than one foot.

Measuring Wire

The Kelvin Test Base makes it easy to measure wire resistance.  Most wire tables show resistance in Ohms per 1,000 feet, this is the identical number as Milli Ohms per foot.  These values range from about 1 Milli Ohm for 10 AWG to 1014 Milli Ohm for 40 AWG all in range on the 4328A for a one foot length.

In Fig 3 above is an HP 4328A, the same model number as my s/n 1828J07728 Milliohmmeter.  But it has a number of differences.

There is no seperate BATT. TEST push button, but that might be done by turning the power push button clockwise (hard to see in photo).
The probe seems to be fixed to the instrument and in this particular eBay auction one of the probes is broken.

Manual Changes Table

The manual is good for the latest version and needs to be back dated for older instrumnets.

Yr

#

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

1967

723/724-00050 to -00120

71

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1967

723/724-00121 to -00210

90

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1967

747/748-00211 to -00270

60

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1968

810/811-00271 to -00300

30

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1968

813/814-00301 to -00341

41

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1968

818/819-00342 to -00440

98

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1968

818/819-00441 to -00730

289

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1970

1030/1040 and below

?

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1972

1210 and below

?

Y

Y

Y

Y

Y

Y

Y

Y

Y

Y

1972

1233 and below

?

Y

Y

Y

Y

Y

Y

Y

Y

Y

1973

1334-02811 and below

?

Y

Y

Y

Y

Y

Y

Y

Y

1974

1420 and below

?

Y

Y

Y

Y

Y

Y

Y

1975

1542J04620 and below

?

Y

Y

Y

Y

Y

Y

1975

1542J04870 and below

?

Y

Y

Y

Y

Y

1975

1542J05050 and below

?

Y

Y

Y

Y

1978

1828J05530 and below

?

Y

Y

Y

1978

1828J06420 and below

?

Y

1978

1828J05711 and above

?

1989 Errata all manuals: C1

1978

1828J07291 and above: C2 this meter is 1828J07728

1978

1828J07961 and above: C3

1978

1828J09594 and above: C4

AC only Power Supply

The stock PCB has two voltage regulator circuits (+13 and -13 VDC) where each regulator circuit uses three transistors (six total)

001 Internal Battery

The only difference is that the plug-in power supply PCB changes.   The option 001 PCB has no active voltage regulation, but just a 15.4 Volt zener diode that feeds the Ni-Cad batteries that are attached to the board through a Silicon diode (0.6 V drop) so open circuit voltage is 14.8 VDC.  Ni-Cad cells are rated 1.2 Volts when out of charge and around 1.4 volts at full charge so a 10 cell pack would provide between 12.0 and 14.0 volts.  The two batteries that make up each pack must be small.

The AC power consumption is specified at 2 watts and option 001 is good for 15 hours.  The batteries, two in each supply (four total) are each 6 Volt 225 mAH.  So for 15 hours operation the current must be less than 15 ma.

Note that when ption -001 is used the Milliohmmeter will only work when the batteries are charged since they are directly connected across the supply rails.  If discharged or bad the meter will not function properly.

For battery only operation the line cord must be unplugged.  The modes of operation are: OFF, Fast Charge (charging but no meter operation), ON charge and meter operation.

The Probe connector is the Hirose RM12BRD-5S and the mating connector is the RM12BPE-5PH or the new number for the lead free version.  I have a few on order from Digikey.
16 Apr 2008 - Digikey shipped the connectors yesterday (the day they were ordered).  Should be here 22 April 2008
22 Apr 2008 - HR1732-ND is the new lead free connector.  The solder cups have an I.D. of 0.040" (measured with gauge pins).  That corresponds to 18 AWG wire but that wire has too much resistance. 

Pin	Internal Color	Function	Connection	Belden 9610
1	Yel	Voltage @ Cur Source	Clip B voltage	Wht
2	Grn	Control of Current Source	Clip A Current	Grn
3	Blk	Current Source	Clip B Current sen	Blk
4	Red	Voltage @ Cur Sink	Clip A Voltage	Red
5	wht/blk	Current Sink - Shield	Clip A Current Shield	Brn Shld

When making a cable there are limitations:
1. for resistance ranges 3 milli Ohms to 100 Ohms there is some maximum resistance and lead length allowed.  The max resistance is determined by the compliance of the Kelvin correction circuitry.  The max length is determined by the accuracy spec and the reactance the lead presents that needs to be removed by the phase detector.
2. for the 1 milli Ohm range the leads must have more than 2 milli Ohms resistance and less than 20 milli Ohms resistance AND the total length must be less than 40 cm (15 3/4").  I think the minimum resistance is related to the current control circuit (pin 2) when the test current is 150 ma.

So to make a cable suitable to use on the 1 milli Ohm range there are limitations on the wire size.

Cable Wiring

See table above.  The A Kelvin clip has Current Sink from pin 5 (as both the Brown wire and at the Kelvin clip the shield around the cable) and connected to the same clip terminal Control of Current Source pin 2 (green cable wire).  The cable shield is not connected at the Hirose connector.  So there's three connections at this clip terminal.  
I've  run the brown and green wires inside the shield braid so you only see the braid in the photos.  One source of error is the Amp Turn product for the area enclosed by the test leads.  If the area changes that can cause an error.  Some experimentation may be needed to see of using two seperate test leads from the connector will work OK.

Cards & Card Sockets

Fig 4 above shows the motherboard which has a socket at the top for the A3 Oscillator card and another socket at the bottom for the A2 amplifier phase detector card.  These are both two row sockets going from 1 to 15 on one side and A to S (A, B, C, D, E, F, H, J, K, L, M, N, P, R, S) on the other side.  The S3 connector has all the pins paired, i.e. 1 to A, 2 to B, etc. but the A2 card only has some pairs jumpered.  The power supply socket that's chassis mounted is a single row A to S socket.

There were stickers all over this instrument, but with a shop rag and some Oops! they are all gone as well as the residue.

The bottom cover can be removed while the feet are installed. 
Using a PosiDrive 2 bit remove the screw and slide the cover back.
Swing the tilt stand bail out to clear the foot and slide cover off.

In order to have an oscillation the probe current source (pin 3) needs to be connected to the current sink (pin 5).  The other pins can be left open and you will still see a little more than a volt rms at Q303 emitter with a frequency near 1 kHz for every switch position.