Windows 7 WUAUSERV With High CPU & No Windows Updates

I have several PCs in my 'lab', a couple of laptops and two desktops. Of these one laptop and the two desktops run Windows 7 Professional. They are all fully licensed.

My main PC is working fine thankfully but both my laptop (a Sony Vaio from around 2011) and my 'spare' PC are suffering from high CPU load from WUAUSERV. This is actually the Windows Update service which provides automatic updating of security updates.

Thankfully both these machines are not turned on much and i have spent almost zero time to figure out what's wrong with them until now when i have tried to rid myself of these Windows Update issues.

My spare desktop used to be my main PC, it's a little old and decrepit it used to run Windows XP but i upgraded due to the EOL issues with XP. It's Windows 7 Professional 32bit.

As it had been my main PC i thought best thing i can do is just wipe it and re-install windows from scratch.

So this is what i did, the only remaining disk in the system was erased and Windows 7 reinstalled from new from the DVD.

After the usual installation of the odd driver i let windows pull down all the necessary windows updates since the creation of my install disk (Win7 SP1), which it did. Some updates installed and some failed. This is in fact normal in the first run of updates. After each reboot it would pick up a few more as dependences and pre-requisites for the updates change as they are installed.

This is where my issues arrived...

After the last run of updates i found the CPU hovering at 50% usage and the updater seems to be frozen checking for updates. The process is SVCHOST.EXE running WUAUSERV as shown in this screenshot:

I have 172 updates now installed on the system and is probably quite up to date right now (April 2016) but it's stuck now so i wont be seeing any new updates going forward. All i get now is 50% CPU load all the time and the Windows Update dialog shows only the following and never completes an update.

Now this issue is not new, indeed Microsoft have a couple of utilities available to diagnose problems like this. These are the Windows Update Troubleshooter and KB947821 aka System Update Readiness Tool (SUR).

The Windows Update Troubleshooter clears the Windows Update cache, logs etc and also checks a number of things that cause issues with updates.

The System Update Readiness Tool does some magic, to be honest i have no idea what it's doing but i think it goes deeper into checking for problems as it's a 200Mb+ installation compared with Windows Update Troubleshooter which is only a few 100Kb.

Both of these were run with no luck, SUR always exits with no reports so i have no idea what it may have looked for, found or tried to fix. The troubleshooter does have a report at the end which details it's unable to fix error 0x800F081F.

Looking around for this error shows it has a plethora of solutions meaning it's probably some generic failure mode without a specific cause. It is interesting though that this was a clean and freshly installed system. My only conclusion is this is caused by one of the updates arriving by the Windows Update mechanism.

I will continue my quest for a resolution and report back when i have some news.

Brandenburg 628-10 10KV Power Supply Teardown

I recently picked up two of these Brandenburg Gamma Series power supplies. A 628-10 (10kV at 10mA) and a 728-10 (20kV at 5mA).

The 628-10 here produced no output but the 728-10 did but has some minor issues. The output HV connector plug was missing, the earth HV return post was broken and was generally a bit grubby.

So i opted to cannibalise the non working 628-10 for it's internal Brandenburg Type 37 HV connectors and a couple of other minor components and scrap the rest.

I expect to be testing and later selling on the 728-10 unit.

DER EE DE-5000 LCR Meter Kelvin Clip Conversion Hack

I recently purchased a DER EE DE-5000 LCR meter, these are well regarded and inexpensive meters made in Taiwan. The unit i purchased was supplied with the TL-21 test leads that plug into the banana sockets and test ports to provide easy to use clips for testing components.

Although the meter is great i was a little disappointed with the test leads, although they are kelvin to the crocodile clip they are very short and dont supply a proper kelvin connection to the component. The two connections are terminated at the solder point on the crocodile clip. So i wanted to change this so i had some proper kelvin probes which had longer wires and easier to use clips.

I was not looking for the ultimate in performance, but just to improve the usability and performance for a reasonable price so i chose to buy a cheap set of kelvin clips off ebay, these cost about £10 including shipping from china. The set i bought was supplied with BNC connectors and coax. This i thought should be ideal for retrofitting into the existing TL-21.

The kelvin clips i bought from ebay.

On opening the DER EE TL-21 test leads, i was pleasantly surprised to find 4 wires running to the crocodile clips and the guard/screen in place which meant converting them with the cheap kelvin clips a simple task.

The BNC connectors were chopped off the new clips and the cable shortened to about 30cm, these were simply soldered in place of the original clips.

After running the calibration (zeroing) procedure on the meter with the new leads, shorting the clips shows 0.00 ohms, it's a shame DER EE dont offer these as an accessory.

Looking At A Cheap L-C Meter Kit From Ebay

In this video i take a look at a cheap (less than £10) L-C meter kit from ebay.

The 'M8' L-C Meter is an inexpensive kit from ebay, it has the following specifications:

Inductance Measurement Range: 0.1μH-2H
Capacitance Measurement Range: 1pF-2.5μF
Electrolytic Capacitor Measurement Range: 0.1μF-30000μF
Assembled Size: 8 * 8 * 1cm / 3.1 * 3.1 * 0.4in
Package Size: 14.5 * 11 * 1.5cm / 5.7 * 4.3 * 0.6in
Package Weight: 85g / 3oz

Reforming Old Electrolytic Capacitors

Recently i had the opportunity to look at reforming of electrolytic capacitors that had been stored unused in a power supply.

The capacitors in question are Cornell Dubilier Computamite branded FAH 42500-30-D3 which are 42,500uF 30V smoothing capacitors from a power supply that was dated March 1973, 42 years old at the time of this blog entry!

The capacitors were stored in-circuit with a bleed resistor of 10 kOhm, storage time was unknown but looking at the rest of the equipment i doubt they would have been used for at least 15 years, probably much longer.

Electrolytic capacitors work by storing a charge on two thin aluminium foil sheets separated by a dielectric material, usually paper impregnated with an electrolyte which is then rolled up into a cylinder and mounted inside an aluminium can with terminals for connection.

During the initial construction of a capacitor one sheet of the aluminium foil is processed to create an oxide layer on it's surface, this sheet becomes the Anode. The oxide layer is an electrical insulation layer between the anode and cathode and is the dielectric, this prevents electron flow and allows the storage of charge. This process is known as forming, the initial voltage is the 'Forming Voltage'. The thickness of the oxide layer depends on the forming voltage, typically the oxide layer is 0.0014uM per Volt so is extremely thin, even on high voltage electrolytic capacitors like large 450v bulk smoothing capacitors the thickness of the oxide layer would only be around 0.0005mm. The greater the thickness of the oxide layer partly determines the maximum voltage that can be applied to the capacitor without breakdown of the dielectric.

If capacitors are left for long periods of time, either in your storage bins, warehouses or even in manufactured products the electrolyte slowly erodes the oxide layer reducing it's insulating properties. If this is allowed to happen only for a couple of years then no issues will arise as the oxide layer will recover when the capacitor is next used. However over longer periods, especially on higher capacity and higher voltage rated capacitors they should be reformed in a controlled manner. Failure to do so may result in localised hot spots between the anode and cathode where vaporisation of the electrolyte can occur which leads to further damage of the aluminium foil.

With this in mind and reading about reforming capacitors in articles it seemed logical to limit the current flow into these capacitors and allow them to slowly build charge to the final voltage and allow the oxide layer to reform in a controlled manner.

Using my Keysight 6632B System Power Supply i initially charged the capacitor to 5V with a current limit of 50 mA. The advantage with these 42,500uF capacitors is their size, it means i can better see the processes involved as the time constant for charging means we have more time to see the process in action.

Initially i noticed that the capacitor certainly drew all the 50mA current available making the power supply operate in constant current mode, the voltage began to climb very slowly to 5V. This indicates the energy from my power supply was flowing into the capacitor but was not being stored as you would expect. I propose this energy was being used to re-establish the oxide layer.

Once the voltage reached 5V the current flow slowly reduced but did not fall to close to zero as i would expect. This indicates that either the capacitor is 'leaky' or as i suspect the energy was being used to reform the oxide layer. The still intact 10 kOhm bleed resistor would account for only 500 uA at 5V.

I disconnected it from the power supply and left it to discharge through the bleed resistor to about 500 mV and charged it again. This time the voltage rose much quicker than previously seen and after it was left connected for several minutes the final current draw was lower too and continuing to drop.

I repeated the procedure again; discharging and charging it but this time i set my power supply to 10V. During charging i noticed that the charge rose normally to 5v but then while charging to from 5V to 10V the voltage rose very very slowly again. I am speculating here that the oxide layer formed initially was enough to block a 5v charge but was not enough to block 10V and so over time the oxide layer increased to slowly block the 10v charge.

I repeated this process in increasing 5V steps, allowing it to take current at the set voltage for about 5 mins at a time until i was able to charge the capacitor to 30.5v and left it connected 'on charge' for several hours.

After finding a somewhat incomplete datasheet for this particular capacitor i went to test the ESR. This i measured using the same method as Alan Wolke (aka W2AEW) detailed in his video #135. After measuring the capacitor for after the treatment i found it's uF and it's ESR was right within specification.

Although these capacitors in question appear to have reformed very well i do not think i would really like to use these long-term given they are over 40 years old but it certainly proves the theory and practice for reforming very old capacitors.

It is vital to limit the current flow into the capacitor during the reforming process. If my understanding is correct this is more critical right at the beginning of the process. So i would recommend limiting the current to no more than 50mA (much less for smaller capacitors) and the voltage to 25% of it's rated voltage and allow it to charge and hold it at that voltage for 10 mins. Then repeat this process in 25% voltage steps until you reach it's operating voltage and then hold it there for 30 minutes.

During the holding time the current flow should drop to almost nothing. After the procedure measure the ESR, leakage and overall capacitance and compare to it's rated specification, if in doubt, discard the capacitor and use a new one.