Mercurial > whatever / changeset

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/content/TeK2445_3.rst	Tue Apr 19 23:15:06 2016 +0200
@@ -0,0 +1,203 @@
+==========================================
+ Tektronics 2445 scope PSU repair: part 3
+==========================================
+
+:Author: David Douard
+:Category: Electronics
+:Tags: Tektronics, 2445, analog scope, repair, test equipment
+:series: Tek2445 PSU repair
+:series_index: 3
+
+
+After having mostly `destroyed the PSU of my Tek 2445
+<{filename}/TeK2445_2.rst>`_, I've decided to be a bit less foolhardy
+and I've subscribed to the `yahoo Tek group`_ and discovered that such
+a failure has been reported there, also after a full recap of the
+preregulator and regulator of the power supply.
+
+I found that I was not so wrong when I built my completely guessed
+LR1060.
+
+It seems to be built on a 75Ω resistor. Mine was not so far, but I
+decided to rebuild one with closer specifications, so I took a new 68Ω
+resistor and I wrapped some 34 gauge wire. The result is a bit
+cleaner, from:
+
+.. image:: {filename}/images/tek2445/lr1060.jpg
+   :alt: first attempt of a hand made LR1060 chocke.
+
+to:
+	 
+.. image:: {filename}/images/tek2445/lr1060_v2.jpg
+   :alt: a better hand made LR1060 chocke.
+	 
+
+But I was stuck because of my dead TR1050. Due to the lack of sucess
+getting the main oscillator starts, I ended by suspecting TR1050
+suffered when Q1050 exploded... And I was right: one of the two
+primary coils was short.
+
+I've found a reasonnably cheap replacement part on `QService
+Electronics`_ shop. It took a bit long to arrive (12 days from Greece
+to France, but to be honest, it seems the longest part of the transit
+was in France), but it arrived. So yesterday evening I've been able to
+replce it.
+
+
+The TR1050 transformer
+======================
+
+This transformer is a small input transformer with 4 coils around a
+common ferrite core:
+
+.. image:: {filename}/images/tek2445/tr1050.jpg
+   :alt: The TR1050 input transformer.
+
+I was not sure of the specifications of each coil, since mine as (at
+least on one coil) defective.  From the schematic, it looked natural
+that both the primary windings should have the same charactritics. So
+when my replacement transformer arrived, I took a few measurements.
+
+The result is:
+
+.. list-table:: TR1050 windings
+   :widths: 15 10 30
+   :header-rows: 1
+
+   * - Winding
+     - R
+     - L
+   * - 1-2
+     - 4Ω
+     - 6mH
+   * - 4-5
+     - 4.5Ω
+     - 6.5mH
+   * - 6-7
+     - 0Ω85
+     - ~ 10µH
+   * - 6-7
+     - 0Ω8
+     - ~ 10µH
+
+I've also quickly checked the voltage ratios using a 42kHz signal:
+
+.. image:: {filename}/images/tek2445/tr1050_ratio_measurement.jpg
+   :alt: Measuring the turn-ratios of the TR1050 transformer.
+
+I use CH2 of my scope to probe the signal produced by the `HP890A4
+<{filename}/hp8904a.rst>`_ linked to pins 1 and 2 (thus a primary
+winding), and CH1 on a secondary winding (pins 6-7), which resulted
+in:
+
+.. image:: {filename}/images/tek2445/DS1Z_QuickPrint17.png
+   :alt: Measuring the turn-ratios of the TR1050 transformer.
+	 
+.. image:: {filename}/images/tek2445/DS1Z_QuickPrint18.png
+   :alt: Measuring the turn-ratios of the TR1050 transformer.
+
+As one can see, the turn ratio seems to be 10. Both the 2 secondary
+windings (5-6 and 7-8) which provide power for the preregulator
+control (U1030) and the inverter drive (U1062, U1064 and U1066), are
+identical to each other.
+
+Testing the new T1050 transformer
+=================================
+
+After having soldered the "new" T1050 transformer, I followed one part
+of the Power Supply Troubleshooting Procedure from the Service
+Manual. In the diagram, there is a path in which there is:
+
+- Unsolder and lift the end of W1060 near pin 7 of T1060;
+
+- Connect the primary test load between the lifted end of W1060 and
+  the sources of Q1060 and Q1070;
+
+- Power up and check for 120V accross load (plus a 1Vpp triangle
+  wave).
+
+  The test load consist in:
+
+.. image:: {filename}/images/tek2445/psu_test_load.png
+   :alt: Primary Test Load.
+
+So I built one and connected it:
+
+.. image:: {filename}/images/tek2445/psu_with_test_load.jpg
+   :alt: The PSU with the test load in place of T1060.
+
+Was a little anxious when I switch the PSU on, but it did not
+exploded. And the voltage across the dummy load was just fine!
+
+So I resoldered W1060, reinserted all the jumpers (between A2 and A3),
+plugged in a dummy load on J303 (unregulated 5V), a tried again. The
+PSU was starting... it was not completely stable though, but I was
+pretty sure this was due to the fact the load was not correct. So I
+reinstalled the PSU in the scope, and gave it a try:
+
+
+.. image:: {filename}/images/tek2445/back_from_the_death.jpg
+   :alt: It's alive!
+
+At last!
+
+
+Summary
+=======
+
+I've spent a lot of time trying to understand and fix this PSU. I
+think I've lost a lot of time and effort because I've been too hasty,
+and I should have taken more time reading the service manuals. By the
+way, the service manual for the Tek 2465A is a better source of
+documentation, for example the question I had about the voltages with
+respect to "REF" on the schematic is answered in this later manual:
+there are indeed 2 references, as I suspected.
+
+The main lesson from this is: never test a switching power supply
+unloaded. I'm pretty sure I blew up the power transistors, the
+transformer and a few other parts because I was switching the
+*unloaded* PSU on and off. It's probably a lesson 101 on switching
+power supplies... never too late to learn.
+
+A the end, I've replaced many parts on my A3 preregulator board. Some
+are probably not required:
+
+- T1050,
+- Q1060 and Q1070: I replaced with IRF820,
+- Q1050: replaced with a IRFP450;the higher gate capacitance seems not
+  to be problematic finally,
+- Q1022, Q1030, Q1040 and Q1063: replaced with 2N4403 and BC337 (the
+  transistors I found at my local components store)
+- almost every capacitor,
+- U1029 and U1040 optocoupler ,
+- U1066: replaced by a DS0026CN bought on ebay
+- CR1040: replaced by an unknown germanium "similar looking" diode,
+- R1060 and R1070,
+- R1069
+
+So I'm done for now with this puppy. It should be completely checked,
+but that will be another story.
+  
+
+Post Scriptum
+=============
+
+For the fun:
+
+.. image:: {filename}/images/tek2445/side_by_side.jpg
+   :alt: Side-by-side: DS1054Z vs. Tek2445
+      
+Despite the fact the Rigol DS1054Z has an incredible number of
+features for which the Tek cannot compete, I still like this
+old puppy a lot. Its front panel is much easier to use: every control is
+directly accessible. And yet, I don't think the Rigol can measure both
+the rise and fall time (with a decent precision) of a square wave:
+
+.. image:: {filename}/images/tek2445/rise_and_fall.jpg
+   :alt: Side-by-side: DS1054Z vs. Tek2445
+
+Beautiful double time-base!
+
+
+.. _`yahoo Tek group`: https://groups.yahoo.com/neo/groups/TekScopes/info
+.. _`QService Electronics`: http://www.qservice.eu/

Binary file content/images/tek2445/DS1Z_QuickPrint17.png has changed

Binary file content/images/tek2445/DS1Z_QuickPrint18.png has changed

Binary file content/images/tek2445/back_from_the_death.jpg has changed

Binary file content/images/tek2445/lr1060.jpg has changed

Binary file content/images/tek2445/lr1060_v2.jpg has changed

Binary file content/images/tek2445/psu_test_load.png has changed

Binary file content/images/tek2445/psu_with_test_load.jpg has changed

Binary file content/images/tek2445/rise_and_fall.jpg has changed

Binary file content/images/tek2445/side_by_side.jpg has changed

Binary file content/images/tek2445/tr1050.jpg has changed

Binary file content/images/tek2445/tr1050_ratio_measurement.jpg has changed