Chemistry Electronics Services
ARCHIVE OF REPAIR TRICKS, USEFUL INFO, ETC.

EDWARDS THERMOCOUPLE VACUUM GAUGE TYPE 507, D35501000 GAUGE TUBE FRIED, NO LONGER SOLD

The gauge-tube filament finally died on an Edwards 507 tabletop vacuum gauge in one of our labs. Sensor part no. D35501000 (actually D355-01-000, also called "TC-1,") with a 5-pin DIN audio connector. Obsolete product, no longer sold by Edwards, and the last gauge tube on eBay was back in 2010. It's a 4-wire type gauge with isolated heater, not the popular 3-wire thermopile Teledyne Hastings DV-6.

At full scale (hard vacuum 10^-5torr) the TC-1 tube puts out the standard 10mV into the 507 meter. The meter's current driving to our (failed) heater filament is 5.00Vdc through a 450ohm resistor and a 200ohm one-turn trimpot set to 80ohms. Heater current is ~10mA for a new tube, or different if the heater wire has gobs of hardened stuff condensed all over it. (NOTE: remove all plumbing and look down into the pipe, and you can inspect the hair-fine wires suspended between the four thicker gold wires.)

     DIN cable pinout:

      1 - n.c.? 
      2 - isolated TC+ (internal rd wire, 10mV out, ~7 to 8 ohms)
      3 - heater (internal bn wire, 10mA drive, 20-50 ohms when new)
      4 - isolated TC- (internal bk wire, gnd return)
      5 - heater (internal gn wire, gnd return)
            Note: any resistance 4 to 5 means internal short, should be >10Meg

Maybe the Hastings DV-23 or KJL-1518 would work as a replacement? Might need a custom paper face for the meter-needle. I have a similar 4-wire tube available: the old Televac "2A" gauge tube, 2-2100-10, from a Fredericks gauge. It's a 10mV fullscale tube, and this one has a 6-ohm heater which needs 0.57Vdc and 89mA. The Edwards 507 gauge only can supply 10mA for the heater. Rats. Even its +12Vdc op amp supply gets seriously dragged down by any extra 90mA. For the heater supply I'll need to add a little AC stepdown transformer, 120 : 6.3VAC, a Zener to give stable AC squarewave volts (two 3.3V zeners back to back in series, with 30ohm series resistor upstream.) Then connect this to the gauge tube pins 1,8 with a 50-ohm trimpot and 22ohm in series. With the sensor pumped down to 10e-4 Torr vacuum, I can set the trimmer to give zero reading on the meter. Works fine, no other modifications needed.

MORE: for aging or contaminated tubes, the filament gets thin and resistance climbs above 50ohms. This causes the meter to drift out of cal at the high-vacuum end, and eventually the blue 1-turn pot for heater-current cannot fix the problem. In this case you can get a bit more life out of the gauge tube by increasing the gain of the op amp. The 11K resistor marked on the board as R6 (one end connected to op amp pin 6) is the feedback resistor. Swap it with an increased value, maybe 13K. That should let the current-adjust pot again reach the vacuum end of the meter scale. Note that if you go too far with this sort of thing, the meter won't read the middle values correctly, even though you've set the 1ATM and 10-3Torr settings correctly.

MORE: lately I've seen two more failed Edwards gauge tubes, but in both cases the heater filament tests good: 40 to 60 ohms (the 7ohm pins, that's the TC.) These tubes were causing the meter needle to slam to full vacuum. It turns out that the heater had become shorted to the thermocouple inside the 4-wire sensor, so it was sourcing many tens of mV through the thermocouple into the front end of the instrument. This is fixable, since the problem wouldn't even exist if the 5V heater-drive supply had been electrically floating. In the 507 meter, unfortunately the heater is powered by the internal +5V supply. Common ground connection.

So, I tap into the 120V connections and add a little stepdown xfrmr etc., with a 5V regulator to provide floating five volts. (Or could have used one of those AC/DC one watt converter modules from digikey.) I slice the pcb traces to disconnect the DIN socket pin-5 heater ground, also slice pcb to remove the trim pot from +5V. Instead wire the heater section to my new floating 5Vdc supply, and yep, works fine. Adjust the heater pot to produce the 10mA, giving 4.50V across the 450 ohm resistor next to the pot. Set meter zero with the front 10-turn pot, then pump it well below 10^3 torr and trim the heater pot to give zero at vacuum.

CALIBRATION:

Remove the two back screws on the bottom. Pop off the gray plastic square ring from the back of the gauge. Pull off the top rear half of the plastic enclosure. That exposes the circuit-board with the small blue 1-turn pot for tube-heater adjust. First set the gauge needle to 1ATM using the front panel 20-turn trim pot screw (turn it all the way up past 1ATM, then back off so the needle just starts to move down.)

Next, I connect and run a turbopump to pump the gauge tube down to well below 1.0 millitorr (at least to 1.33x10^-4 mBar on our Penning gauge.) Then I use the internal blue pot (heater current adj.) to set the gauge needle so it points to a spot that's 2.0mm below the lowest milliTorr tick mark (2 mm below the 1x10-3mT mark.)[*] This seems to be right on. When the turbo is spinning down and the pressure rises to 1.0 millitorr, the needle is pointing to the right place.

[*] If this pot cannot bring the needle down far enough at the high-vacuum end of the range, then your gauge tube has aged too much. Find a new one (see above.) Or, if the failure remains small, then do as above, and swap the 11K R6 resistor with a larger value, 13K or 16K.