content/hp34970a.rst

changeset 58
58e79464cd09
child 115
6b6e13653348
equal deleted inserted replaced
56:3d0bed1a79c8 58:58e79464cd09
1 =================================
2 HP 34970A Data Acquisition Unit
3 =================================
4
5 :Author: David Douard
6 :Category: Electronics
7 :Tags: HP, 34970A, DMM, repair, test equipment
8 :series: HP 34970A repair
9 :series_index: 1
10
11
12 I recently "scored" a HP 34970A with the DMM included and 2 plugin
13 cards (HP 34901A and HP 34907A) for less than 200€. The plugins should
14 be in working condition, but the main switch unit was almost "given"
15 with the plugin modules since it is non functional.
16
17 .. image:: {filename}/images/hp34970a/front-view.jpg
18 :alt: front view of the HP 34970A data acquisition unit
19
20 It's in very good shape, the front panel and the push buttons are
21 clean. The knob is nice but it has a rough step when turning it for a
22 complete revolution.
23
24 When powered up, however, things are a bit nastier... The VFD is, as
25 often with these equipment, very dim and the digits tend to puke on
26 the others. In fact, it's mostly unreadable, and the unit beeps
27 several times, signs of errors (which I cannot really read them see on
28 the display).
29
30 Internals
31 =========
32
33 The unit consists in 4 parts:
34
35 - the main logic/cpu board on the bottom left side,
36 - the DMM module, juste above the main board,
37 - the frontpanel and
38 - the back plane where plugin modules are connected.
39
40 .. image:: {filename}/images/hp34970a/bottom-view.jpg
41 :alt: bottom view of the HP 34970A
42
43 .. image:: {filename}/images/hp34970a/top-view.jpg
44 :alt: top view of the HP 34970A with DMM removed
45
46 .. image:: {filename}/images/hp34970a/dmm.jpg
47 :alt: DMM module of the HP 34970A
48
49 The front panel hold the keyboard, the rotary encoder and the diplay system.
50
51 .. image:: {filename}/images/hp34970a/front-panel.jpg
52 :alt: front panel of the HP 34970A
53
54 The front panel is managed by a 80C51 (a 87C51 actually)
55 microcontroller. It's responsible for handling the keyboard and
56 feeding the shift registers of the VFD drivers (a pair of
57 HV518PJ). The communication with the main CPU is a serial line with
58 opamps to adjust the signal levels.
59
60 The chips are PLCC44 located under the VFD are the drivers:
61
62 .. image:: {filename}/images/hp34970a/front-panel-pcb.jpg
63 :alt: front panel PCB of the HP 34970A
64
65 On my unit, the voltage levels for the VFD are fine. I suspected that
66 one of the 2 VFD drivers (the one responsible for driving the grids)
67 were partially failing, so I ordered a pair of new drivers, as well as
68 some PLCC sockets. It was my first real smd revork. I though that it
69 might be easier to cut the pins of the soldered chip, but it was not a
70 good idea, I peeled 2 pads on the PCB. Thanksfully, they were not
71 broken. So I've undoldered the other VFD with my new cheap hot air
72 soldering station, and it's been way easier and did not damage the PCB
73 at all... Lesson learned.
74
75 After that first real SMD unsoldering experiment came the first SMD
76 soldering: install the plastic made PLCC sockets... I was not sure
77 wether to use the hot air gun or my Weller soldering iron, nor what a
78 correct temperature for the hot air station whould be adequate. The
79 informations on the socket datasheet were not very clear to me.
80
81 So I gave a try (I think at around 280°C). The sockets seemed properly
82 soldered, but in fact, a few pins were not making contact. So I
83 decided to cut the bottom of the socket (which is essentially useless)
84 so I can more easily rework these "cold" pins with my soldering iron.
85
86 The tricky part of installing these PLCC sockets is that decoupling
87 capacitors (C6, C7, C9 and C10) are a bit too close, so I had to move
88 them.
89
90 .. image:: {filename}/images/hp34970a/front-panel-sockets.jpg
91 :alt: PLCC sockets for the VFD drivers of the HP 34970A
92
93
94 Now I had to try to rejuvenate the tungsten filaments. Like for old
95 CRT (and even for triodes), the filament, which is responsible for
96 emiting electrons toward the anode (coated with phosphor) tend to be
97 polluted and loose their emissive power. A common trick to "clean"
98 them is to apply a much higher voltage than rated so they shine for a
99 few seconds. One mst be careful not to burn them, otherwise the VFD is
100 definitively lost.
101
102 .. image:: {filename}/images/hp34970a/vfd-back.jpg
103 :alt: back side of the VFD (probably dead) module
104
105 So I tried such a rejuvenation on my VFD module. I wasn't sure what
106 color the filaments must be heated to (between a light orange to an
107 almost white yellow). I was doing this using my `Lambda PSU
108 <{filename}/LPD422FM.rst>`_ raising the voltage while looking at the
109 filaments.
110
111 The results are very disappointing. The VFD is a little bit brighter,
112 but the spilling over between digits is now much worse...
113
114 I guess this VFD module cannot be saved, after all...
115
116 So I've been thinking of a plan B, and I think I'll try to replace the
117 VFD display with a blue 256x32 OLED module. It's smaller (I cannot
118 find a module near the size of the VFD display), but it should still
119 be quite readable. Indeed, that requires a bit of work: I plan to
120 build an arduino-based adapter (using an arduino micor) that will take
121 place of the VFD drivers (then be fed by the serial-to-shift registers
122 outputs of the 80C51). One other solution is to replace the 80C51 with
123 a newly programmed one, but I think reverse ingineering the serial
124 communication between the front-panel microcontroller and the main CPU
125 would require much more work. Since it's a 87C51, I doubt I can easily
126 dump it's ROM...
127
128
129 Meanwhile, next step is to take care of all these errors.

mercurial