comparison: content/HPZR24W.rst
content/HPZR24W.rst
- changeset 72
- 9296d8c2339a
- child 115
- 6b6e13653348
equal
deleted
inserted
replaced
| 71:76d5c4108e51 | 72:9296d8c2339a |
|---|---|
| 1 ===================== | |
| 2 HP ZR24W PSU Repair | |
| 3 ===================== | |
| 4 | |
| 5 :Author: David Douard | |
| 6 :Category: Electronics | |
| 7 :Tags: ZR24W, PSU, PWB-1336-02, TPS-10036 | |
| 8 | |
| 9 A friend of mine had a problem with his (a bit old) 24" HP monitor: the device | |
| 10 was working properly, but only for a while (few minutes at most). | |
| 11 | |
| 12 Looked like a good candidate for a quick repair, problem being most probably | |
| 13 bad caps are so. | |
| 14 | |
| 15 The first step was to find out how to disassemble the down thing, since these | |
| 16 "modern" things are not designed to be serviced and have no screw, only plastic | |
| 17 clips which are so easy to break. | |
| 18 | |
| 19 Whatever, I finally succeded in opening the box (with only one or 2 broken | |
| 20 clips). | |
| 21 | |
| 22 Overview | |
| 23 ======== | |
| 24 | |
| 25 Inside the beast, the usual stuff under the shields: | |
| 26 | |
| 27 - a PSU, | |
| 28 - a main controller board, | |
| 29 - an LCD driver board, and the monitor being a bit old, | |
| 30 - a blacklight HV driver board. | |
| 31 | |
| 32 | |
| 33 .. image:: {filename}images/ZR24W/back.jpg | |
| 34 :alt: Back of the HP ZR24W | |
| 35 | |
| 36 First thing, to make sure the problem comes from the PSU, I powered the monitor | |
| 37 from my bench power supplies. I used one output of my HP E3648A to generate the | |
| 38 19V rail that powers the backlight. It was a bit short (CC limited at 16V), but | |
| 39 enough to power the backlight. The second output was used to produce the 12V, | |
| 40 and the 5V was generated by my very old home made PSU. | |
| 41 | |
| 42 .. image:: {filename}images/ZR24W/ext_power.jpg | |
| 43 :alt: Testing the display with my bench PSU. | |
| 44 | |
| 45 The monitor was working fine when powered from these PSUs, so the problem was, | |
| 46 indeed, this Tatung PWB-1336-02 switching PSU: | |
| 47 | |
| 48 .. image:: {filename}images/ZR24W/psu_top.jpg | |
| 49 :alt: Top of the Tatung PWB-1336-02 PSU | |
| 50 | |
| 51 .. image:: {filename}images/ZR24W/psu_bottom.jpg | |
| 52 :alt: Bottom view of the Tatung PWB-1336-02 PSU | |
| 53 | |
| 54 The design is nice and clean. Electrolytic caps however are not the best ones | |
| 55 (mostly Lelong ones I think), which is not really a surprise in this kind of | |
| 56 device. | |
| 57 | |
| 58 At first sight, nothing strike the eye, no leaky cap, no burnt resistor or | |
| 59 PCB... | |
| 60 | |
| 61 Not a 10mn fix, in the end. | |
| 62 | |
| 63 The PSU generates 3 voltages: | |
| 64 | |
| 65 - 5V @2.7A | |
| 66 - 12V @0.8A | |
| 67 - 19V @3.2A | |
| 68 | |
| 69 | |
| 70 Finding the problematic power rail | |
| 71 ================================== | |
| 72 | |
| 73 There is a hard switch on the PSU, next to the IEC socket, and When the power | |
| 74 is on, the 5V is hot, wether the monitor is on or in standby. | |
| 75 | |
| 76 When the display is powered on, the 2 other voltages are started. | |
| 77 | |
| 78 The pin 10 on the cable between the PSU and the main CPU board is dedicated | |
| 79 to the "power saving" state. It must be high (at 5V, which is always present) | |
| 80 to enable the 2 other voltages. | |
| 81 | |
| 82 Note that there are 2 other pins dedicated to power management (pin 11 and 12, | |
| 83 marked as "On/Off" and "Vadj"). But these are directly routed to the backlight | |
| 84 board and take no part in the PSU management. | |
| 85 | |
| 86 | |
| 87 My first test has been to plug my cheap `electronic load | |
| 88 <{filename}/ZPB30A1.rst>`_ on the 5V with the 2 other voltages stopped. | |
| 89 | |
| 90 And I could reliably sink 3A from there. So the problem must be on one of the 2 | |
| 91 other rails. | |
| 92 | |
| 93 But I also could sink the max amperage from the 2 other power rails (1A from | |
| 94 the 12V and 3.5A from the 19V)... | |
| 95 | |
| 96 Ok, so each power rail seems to work fine alone. | |
| 97 | |
| 98 But when I sink current from the 5V rail while the 2 other voltages are up, | |
| 99 then the PSU fails after a short while. | |
| 100 | |
| 101 I've tried to probe a bit the switching curves using my Rigol DS1054, but the | |
| 102 PSU being "hot" (around 400V), and having no isolation transformer, I could not | |
| 103 probe the signals correctly (using 2 probe and displaying the A-B curve, which | |
| 104 is some kind of a joke on the Rigol, since you cannot hide the A and B curves: | |
| 105 the curve substraction is purely computed from displayed curve. When I found | |
| 106 how useless this was on the Rigol, I had not enough energy to extract my old | |
| 107 `Tek 2445A <{filename}/TeK2445.rst>`_ from under the pile of test equipment | |
| 108 stowed in the closet...) | |
| 109 | |
| 110 | |
| 111 A glimpse at schematics | |
| 112 ----------------------- | |
| 113 | |
| 114 At this point, I needed to try to understand a bit the schematic. I wasn't even | |
| 115 sure the problem came from this side of the PSU, it could also be a problem in | |
| 116 the "isolated" part (bad caps, bad optocoupler, bad voltage reference...). I | |
| 117 had checked the main caps, and they seemed to be ok (not quality japanese | |
| 118 brands, but still the correct value and low ESR). | |
| 119 | |
| 120 | |
| 121 On the live part, before the transformers, there are 2 chips : | |
| 122 | |
| 123 - a TNY279PN_ dedicated to the 5V rail | |
| 124 - a CM6807_ for the 19V rail | |
| 125 | |
| 126 The 12V is produced from the 19V rail by a small DC to DC converter (FP6185_). | |
| 127 | |
| 128 In fact, this PSU design is almost just the 2 application circuits (found in | |
| 129 their respective datasheets) merged together. | |
| 130 | |
| 131 The noticable points are: | |
| 132 | |
| 133 - The TNY279PN DC input can be come from either the input bridge rectifiers | |
| 134 (via a diode and a thermistor), or, when powered, by the DC produced by the | |
| 135 CM6807 via the PFC circuit. So when the CM6807 is down (when the monitor is | |
| 136 in standby), the DC input for the TNY279 is around 318V, but when the monitor | |
| 137 is on, the CM6807's PFC circuit rise this voltage to almost 400V. | |
| 138 | |
| 139 - The bias winding, normally used to power the TNY279 (to allow a very low | |
| 140 no-load power consumption, datasheet says <50mw) also powers the CM6807. | |
| 141 | |
| 142 A funny side effect of this design is that the CM6807 cannot be started if | |
| 143 there is no load on the 5V rail: in this case, the auxiliary voltage of the | |
| 144 transformer managed by the TNY279 is not enough to start the CM6807. | |
| 145 | |
| 146 | |
| 147 The application circuits are as follow: | |
| 148 | |
| 149 .. image:: {filename}images/ZR24W/tny279_app_circuit.svg | |
| 150 :alt: Example application circuit for the TNY279PN controller (from the `datasheet <TNY279PN>`_) | |
| 151 | |
| 152 .. image:: {filename}images/ZR24W/CM6807_app_circuit.svg | |
| 153 :alt: Example application circuit for the CM6807 controller (from the `datasheet <CM6807>`_) | |
| 154 | |
| 155 In this PSU, the input DC of the TNY279 (the point just before R5 in the app | |
| 156 circuit) in connected just behind the PFC circuit, i.e. just after the D2 diode | |
| 157 of the CM6807 app circuit. | |
| 158 | |
| 159 In order for the PSU to generate the 5V rails even when the CM6807 is off, | |
| 160 there is derivated DC input path, from the bridge rectifier to the input DC | |
| 161 rail, consisting in a diode (D922 on the PCB) followed by a thermistor (R915): | |
| 162 | |
| 163 .. image:: {filename}images/ZR24W/input_DC.jpeg | |
| 164 :alt: Input DC showing both paths (yellow: via the PFC, red: the alternate | |
| 165 path when power saving is on). | |
| 166 | |
| 167 | |
| 168 Finally, the culprit | |
| 169 -------------------- | |
| 170 | |
| 171 While testing the PSU powering the CM6807 from an external source, I notices | |
| 172 another strange behaviour: the TNY279 enters a failure as soon as I sink | |
| 173 current from the 5V rail, but it remains faulty as long as I let the CM6807, | |
| 174 thus the PFC running, and the input DC at 400V (instead of 318V when the PFC is | |
| 175 not activated).. | |
| 176 | |
| 177 At this point, I though there is not way the problem can come from the 5V | |
| 178 regulation loop, neither from the CM6807 circuit, so the only culprit left | |
| 179 would be either the TNY279PN or a component close to it, a cap (especially the | |
| 180 BP/M one), a diode or a resistor. So I tested these parts (again, to be fair), | |
| 181 and I replaced the cap connected to the BP/M pin. With no improvement, the only | |
| 182 remaining faulty part must be the TNY279PN. | |
| 183 | |
| 184 So I decided to buy a couple of them from RS, and 2 days later I replaced it, | |
| 185 which did fix the PSU. | |
| 186 | |
| 187 What a tricky half-failing part! | |
| 188 | |
| 189 | |
| 190 | |
| 191 | |
| 192 .. _TNY279PN: | |
| 193 https://www.power.com/sites/default/files/product-docs/tny274-280.pdf | |
| 194 .. _CM6807: http://www.championmicro.com.tw/datasheet/Analog%20Device/CM6807.pdf | |
| 195 .. _FP6185: http://www.feeling-tech.com.tw/km-master/ezcatfiles/cust/img/img/24/fp6185drv02-g1.pdf |
