Est. 1984

This kit provides the most common 'fix' for the well-known capacitor problem that plagues the old power supplies in the BBC Micro model A/B, B+ and Master 128. The original X2 capacitors degrade over time and often fail quite spectacularly, leaving behind a ruptured capacitor and a rather pungent smell. It is better to change the capacitors before they fail rather than after.

For £2.25 (+P&P) you get 10nF and 100nF RIFA Kemet X2 suppression capacitors (both rated 110 degrees), a Rubycon 220uF 35V 105 degrees electrolytic, more cable ties than you are likely to need and a replacement strain relief bush for the incoming mains cable. At the time of writing it's one of the best prices around and yet offers more than any other Model B capacitor kit.

These capacitor kits are normally classified as a 'large letter' by Royal Mail. This kit will be sent by Royal Mail 2nd class and there is no CD included.

This kit contains the three main devices required for the most common repair or preventive service. These are a 10nF and 100nF X2 rated mains suppression device and a 220uF electrolytic capacitor. These are not budget components. The ones shown are the ones that will be supplied, are sourced from a reputable UK or global supplier and have the correct lead spacing to match the holes in the power supply PCB.

Removing the power supply circuit board from its metal frame involves cutting a few cable ties. The precise number varies depending on the machine. When the power supply is reassembled, you need to (ideally) replace these ties. Suitable cable ties are supplied as standard and there will be more than enough.

Servicing the power supply involves working on mains equipment. Capacitors can retain a charge long after the mains supply has been removed and you should only attempt a repair of the power supply if you have the necessary experience and tools.

Capacitor C9 - Any advantage in a +125 degree Celsius rating?

Before answering this question, it's useful to have a realistic idea of how hot the electrolytic capacitor C9 inside the power supply actually gets. To establish this, I dismantled the power supply and attached a temperature sensor to C9. For safety reasons I don't recommend you do this at home. I've performed this experiment so you won't have to.

The machine was a Model B with an issue 7 board. Add-ons included a 1772 DFS upgrade, a ROM/RAM expansion board providing the full 16 sideways ROM sockets supported by the Beeb, a micro SD card in the User Port and a second processor attached to the 'Tube'.

On a flat surface (ie ventilation slots clear) the temperature in the PSU eventually stabilised at 67 degrees. Allowing for a slightly warmer day (ambient temperature), this could rise to 75 degrees. This is well short of the 105 degree rating of many capacitors.

As an example for comparison purposes, consider the Wurth Electronik RedExpert 220uF 25V electrolytic capacitor rated at 125 degrees. One obvious difference is that the voltage rating is only 25V compared with the 35V for the Rubycon part. The measured ESR for the Rubycon was slightly lower than the Wurth component, although not significantly so. (ESR = Equivalent Series Resistance. Lower is better)

The most interesting point comes from the expected life of the device. For the Wurth capacitor, the datasheet quotes 2000 hours @ 125 degrees. For the Rubycon it is 7000 hours @ 105 degrees. That's over three times longer.

However, data sheets typically quote average lifetimes at a certain temperature. There are some rules of thumb (and arcane equations) which can be used to estimate the capacitor life at different temperatures. Applying such rules for operation at +75 degrees you get a predicted life for both the Wurth and Rubycon of over 50,000 hours. Based on 3 hours use per day, it works out at around 50 years. However, other factors (such as degrading over time) then play a part. Indeed, some manufacturers make it clear that any predicted lifespan in excess of 15 years should not be relied upon.

The Wurth device is a good quality part but there is no evidence that it will last longer than the Rubycon device. When comparing products, don't be misled by unsubstantiated claims that product A will last longer than product B.

Is the 125 deg. C Wurth capacitor new?

Not really. It's certainly been available for nearly six years and possibly much longer. There's nothing particularly new about it.

Is the Wurth capacitor more expensive?

More expensive than what? Comparative ads can be permitted, but they need to compare similar products serving a similar purpose. A comparison with nothing is meaningless.

The Wurth and Rubycon capacitors are actually very similar in price. Indeed, you can select the Wurth or Rubycon device at random and then 'prove' that it is more expensive than the other. In fact, it's even possible to show that they cost exactly the same. You can do all of these things just by carefully choosing the quantity and/or supplier.

So, it is misleading to state that one device costs more than the other. Even when there is a price difference, it can be as little as 0.1p. There is no plausible argument that one item is better than another just because under contrived circumstances it costs 0.1p more. Quite apart from that, advertising guidelines do not allow objective claims of better reliability based purely on price.

The +125 degrees rating, fancy name and red case do not make the Wurth superior. When used in the Beeb it has no substantiated advantages and, as shown above, any suggestion that it's more expensive is misleading.

The parts supplied in this kit are;

Suitable soldering iron

Compared to the Beeb's motherboard, the power supply's circuit board has relatively large areas of copper, thicker component leads and more solder on the joints. To melt the joints you will need a bit which has a reasonable thermal mass backed up by an iron with suitable wattage. A 1mm bit attached to a 15W iron will not be suitable. A 3mm bit on a 50W temperature controlled iron and you're good to go.

The relevance of this, apart from the obvious, is that some capacitor data sheets make the point that prolonged heating can cause damage and soldering a lead should take no more than about three seconds. This is actually easier to achieve with a suitable iron/bit combination.