On Site Time today was spent on the T86 with cleaning, corrosion treatment and some painting. The plastic corrugated cable hoses on the aerial assembly have now been cleaned with sugar soap and rinsed. A task needed as dirt was filling up the corrugations in places!
An inspection/alignment cover was removed in the base of the pedestal and the accompanying photo shows an exposed gear wheel covered in a granular form of corrosion (rusticles?).
The granules can be scrapped to loosen them and then vacuumed out carefully avoiding the oil bath. Only this one section of the gear wheel was visible but once the team are back on site the aerials will be moved to check and remove any further corrosion granules.
The base of the pedestal continues to be cleaned and red oxide painted.
Away from Cosford We have an unserviceable CHARGE Pixel Store, a spare is available so the simulator is not affected. To fault find a Pixel Store a test rig is set up to use one using a serviceable CHARGE card to drive one display.
We do not have any test software or routines to test CHARGE cards so the approach is to set up the test rig and create a set of reference waveforms (screen shots from a USB driven scope) at key points in the card’s logic We then swap the U/S card and look for differences. This process is usually reiterated a few times but has been successful in fixing two Pixel Stores in the past. With a U/S Pixel Store fitted the monitor display has lines laid over the firmware display as seen below .
The firmware display appears in the LCP prior to the Bloodhound software booting and is generated by CHARGE, not the Argus 700 computer.
Before running up the simulator there were two jobs to carry out. Firstly, replace a CHARGE Controller that had a low resistance AVX glass encapsulated decoupling capacitor and secondly fit a monitor back on its runners and in to the display console, a two man (in our case) job.
Here is the CHARGE controller that had been repaired due to low resistance AVX capacitors, now a common problem. The photo shows two original AVX glass decoupling capacitors below the ICs and two yellow replacements fitted during repair off site using a test configuration in BMPG’s ‘third line depot’.
Running up gave no errors initially but one display (instructors) then looked to have lost a colour. Investigation revealed a bad connection between a CHARGE Controller’s DIN connectors and the DIN socket on the backplane. A bit of reseating sorted that problem. Historically the CHARGE chassis and connectors have taken a hammering after years of being abandoned in the open, often with the LCP door open with the CHARGE box taking the full force of the elements.
A second run of the simulator at the end of the day was problem free. An engagement was run, all serviceable with no problems.
The main effort as painting the aerial assembly of the T86. Most of the painting of the aerial beam and mass of waveguides has been completed.
Also, most of the inside of the pedestal repainted with red primer, as per the original.
The T86 is now looking substantially better from when first recovered. Here you can see the general work environment for working on the T86 aerial assembly and cabin roof.
What a great day weather wise. The first time the hangar doors have been opened and no blast of cold and damp air. It was actually warm in the hangar which means – painting.
Before getting down to the painting there was some unfinished business with the CHARGE cards to ensure all four display channels had a good and proven set of cards. The problem with all the I/O cards, included CHARGE, is the glass encapsulated AVX decoupling capacitors. A simple ‘resistance between +5V and 0V’ exposes the problem and one CHARGE controller card was found to have 8R when the expected is around 250R. Unfortunately, or fortunately, the 120 amp +5V supply compensates but eventually it will mean problems – a Digital Output card has already been damaged with a U/S (low resistance) AVX capacitor trying to burn a hole though its PCB! The low resistance Controller has been removed and taken off site for repair by what is ex RAF 3rd Line supported by ex-Ferranti engineers.
The simulator was given two good runs during the day and apart from the Floating Point card which didn’t reset and was replaced, the simulator is running well. With everything serviceable the simulator boots in around five seconds. The one second timer is running before the displays have warmed up! The fast boot is down to replacing the original Winchester disk with a solid state emulation – the LCP now runs off an SD card!!!
The majority of time during the day was taken up with a job which everyone has been putting off – painting the wheels from the T86.
At least it was pleasant painting outside the hangar in the warm. With the simulator running well the focus now returns to finishing the restoration of the T86.
Apologies for delay in posting Pete’s blog for last Saturday, domestic jobs got in the way! Mike S