Posts by John Dawson

Famously flawed, it is 30 years since the Hubble Space Telescope was launched

Re: Mirror

Are you Steve Gull? If so I claim my £5.

MPs to grill Post Office and Fujitsu execs on Horizon IT scandal after workers jailed over accounting errors

File on Four podcast

https://www.bbc.co.uk/programmes/m000f5hb

This covered the topic rather nicely some 3 weeks ago. I recall there was testimony from Fujitsu insiders which iirc stated that they could remotely access the accounts at any time, without the post office counter staff knowing, and that this might well cause errors if transactions were taking place at the same time. This was contrary to evidence given in court that no-one, apart from the user, had access to the account.

Well worth a listen.

Apple: EU can't make us use your stinking common charging standard

Re: Gold plated cables

Not attacking your point of view but you can make a pretty good single ended amplifier with split power rails and an active current source for the non-output side. So no big electrolytics in series with the signal. And the miracle of negative feedback (as well as proper supply decoupling for the driver stages) can eliminate power supply sensitivity for all practical purposes.

PS - I design this sort of stuff (though not precisely what I've described above) for a living.

PPS - agree with you on the mains power supply cables - any effects are probably mostly related to RF suppression, or lack thereof.

Alphabet, Apple, Dell, Tesla, Microsoft exploit child labor to mine cobalt for batteries, human-rights warriors claim

Re: One-eyed selectivity

To the best of my knowledge no cobalt is used in silicon based solar panels, which account for 90% or more of solar cells. It may be in any associated lithium battery storage but there's not a lot of that. Finally that may change in the future with other types of bulk battery storage (flow batteries for example).

Stand back, we're going in: The Register rips a 7th-gen ThinkPad X1 Carbon apart. Literally

Re: Who actually needs the talking to?

Aah the T600 - I am still running one on Win 98 to support an Audio Precision analyser. It's been almost daily use for the past 15 years!

They don't make 'em quite like they used to....

'Literally a paperweight': Bose users fume at firmware update that 'doesn't fix issues'

Re: How did we get here exactly? HDMI strikes again

Aaah HDMI - we used to call it Hardly Defined Mostly Intermittent back in the day..... it's still tricky to implement 15 years later and if you choose the wrong part as the equipment manufacturer you may just be stuck. I have a (small) amount of sympathy for Bose in this instance.

Gee, SEC, how did that get out?! 'Leaked' Tesla email claims big boost in Model 3 production

Re: Wrong

You are not very observant then - I see them every day here in Cambridge. Mostly Model S.

Tesla big cheese Elon Musk warns staffers to tighten their belts in bid to cut expenses (again)

Re: Servicing

Actually on a Tesla there is minimal brake wear because of regenerative braking. You can drive the car almost entirely on one pedal.

Tesla's chief accounting officer drives off after just a month on the job

Re: The product is strong, and the vision is excellent

For the record my Tesla Model S costs about 2.4p per mile to run using overnight economy 7 electricity; I have calculated that you can charge about 10M vehicles this way in the UK before the grid reaches capacity. About half my car's electricity comes (indirectly of course) from the modest solar array on my roof. That will take us perhaps 10 years forward giving time to work out improvements in electricity generation, storage and distribution.

Taxing vehicle usage is another matter entirely but it clearly has to change and that may alter my present economic model :-(

Q: How do you test future driverless car tech? A: Slurp a ton of real-world driving data

Tesla's advantage

An interesting experiment - Tesla have been doing much of this for several years with a fleet now approaching 200K cars - any that are equipped with autopilot HW. Every car has an always on 3G/4G connection back to the mothership. Of course what they are doing with all that data is the big question.

SpaceX sneaks in X-37B space plane launch ahead of Hurricane Irma

Iain M Banks lives!

"Of Course I Still Love You" - wonderful :-)

Smart meter firm EDMI asked UK for £7m to change a single component

Re: Solar Panels

Here is a simplified explanation for the UK. It turns out that most charging (approaching 90%) of electric cars is done at night, so for a very long time to come you can use the gap between day and night demand to charge these vehicles. This yields about 10-12GW of capacity over about 7-8 hours, conservatively 80GWh per day. At 250Wh per mile that allows you to charge say 12 million electric vehicles without needing extra grid capacity. The full UK car fleet is about 32 million vehicles, averaging 7,900 miles per year by the way. Of course you will need to burn extra fossil fuel but that is more than offset by the diesel and petrol you no longer need. More to the point this will clean up our city air considerably.

Kotkin: Why Trump won

This guy Jonathan Pie really gets it

I have been out all day so missed this thread until now. I ploughed all the way through and was about to post the same video, which appears to have gone viral (now around 9 million views) and appeared on my Facebook account this morning. You beat me to it Matt 4 :)

Anyone still following this thread PLEASE take a look at the video - it sums up so much so eloquently in only 6 minutes! Afterwards some of you may wish to look closely at yourself in the mirror.

Brit AI genius dies

Shocked

It's not often I get all worked up about the untimely death of someone I barely knew but on reading this I am in a state of deep shock. I knew little of David's information theory background but I heard him lecture a couple of times in Cambridge on the topic of renewable energy and not only bought the book for myself but for my children too.

His no-nonsense approach means he will be sorely missed by not only his family and friends but in influencing the government too. He was the kind of scientist we'd all like to be.

My sincerest condolences to you family David and RIP.

Innovation creates instability, you say? BLASPHEMY, you SCUM

Re: Long live the BS meter

I read that as smelly our mum - oh well....The card however is priceless - thanks, it has made my day! I guess I am easily pleased :)

Fukushima fearmongers: It's your fault Japan dumped CO2 targets

Another troll surely? OK, on reflection, it's just more misinformation.

Once again I refer you to this paper which tells us the radiation for naturally occurring potassium 40 in any fish is about 15x higher than the Cs contaminated fish caught off the coast of the USA. More interestingly the radiation from the naturally occurring Polonium in the fish is some 500x the Cs dose.

http://www.pnas.org/content/early/2013/05/30/1221834110.full.pdf+html

Re: Anti-nukers... vs. Pro-Nukers

With regard to the risk from Pacific fish irradiated from Fukashima it would appear that the additional risk is effectively infinitesimal according to this paper - http://www.pnas.org/content/early/2013/05/30/1221834110.full.pdf

Basically one normal banana would give you about 20x the 137Cs radiation dose in 200 g of Fukashima-contaminated fish, from the naturally occurring Potassium 40 in the banana!

Here's the meat so to speak :

Dose to Humans. Consumption of 200 g (a typical restaurant-sized

serving) of PBFT contaminated with 4.0 Bq·kg-1 dry weight of

134Cs and 6.3 Bq·kg-1 dry weight of 137Cs (mean values for PBFT

caught off San Diego in August 2011) resulted in committed

effective doses of 3.7 and 4.0 nSv, respectively (Table 1). To put

this into perspective, the combined dose of 7.7 nSv from these

two Cs isotopes is only about 5% of the dose acquired from

eating one uncontaminated banana (assuming 200 g weight) and

absorbing its naturally occurring 40K (28), and only about 7% of

the dose attributable to the 40K in the PBFT (Table 1). More

strikingly, the dose from both Cs isotopes is only 0.2% of that

attributable to the naturally occurring 210Po from ingesting the

fish (Table 1). Furthermore, in August 2012, PBFT off California

were found to have less than half the levels of radioactive Cs

than were found in August 2011 (29), which would result in even

lower doses to human consumers.

HTH!

Skyfall makers 3D printed Bond's DB5

Cinema pricing can be OK

Saw the movie last night at my local Picturehouse in Cambridge with digital projector and a THX certified sound system. Very impressive and my partner, who is not a big JB fan,really enjoyed it too.

With membership and Orange Wednesdays concession it cost less than 7 GBP for two tickets. Yay!

John Dawson

Apple: I love to hate, and hate to love thee

Lenovo

Lenovo offer a 3 year carry-in warranty on their T series machines - I swear by these and have owned them back from the IBM days. The warranty is transferable too, so if you buy one a year old on e-bay you still have two years left. So far I have never had a machine fail in the three year period anyway.

IBM/Lenovo also have comprehensive service manuals on-line so you can fix most faults yourself outside warranty.

Works for me :)

Arcam rPac

Re: jitter-free analogue signal

Umm... to which TI DAC are you referring? It is certainly not the PCM5102. Perhaps you are thinking of the PCM270x family? These parts are limited to 16 bits 48ks/s and have a crude clock recovery system that produces LOTS of jitter related products in the audio output. They are fine for simple USB speakers etc but are not hi-fi in the highest quality sense of the word.

Please read my main posts in this thread to find out more.

John Dawson

Re: jitter-free analogue signal

Jitter is indeed about small timing errors in digital signals, below the amount needed to cause real bit-errors.

This would not normally matter except that in digital audio it is necessary to have a really clean clock at the audio DAC (or ADC); otherwise errors in the timing are equivalent to linearity errors in output level from one bit number to the next.

The problem with digital audio interfaces (S/PDIF, HDMI and USB for example) is that no clean master audio clock is actually transmitted over the interface - you have to reconstruct it at the receiving end. This is not as easy as it sounds. One way is to use a FIFO and clock out using a clean clock and ask the transmitter to vary its rate so as to avoid overflow/underflow. This is impossible with S/PDIF as it is one way but is possible with USB provided you use the method. There is not much silicon around that actually does this - Arcam is one of the few companies to implement it in all its USB DACs right down to this inexpensive one. The other solutions use the 1millisecond frame rate associated with USB and this produces a LOT of timing errors i.e. jitter which modulates the audio output signal.

As to what it sounds like - well that depends on both the amount and frequency spectrum of the jitter. Where it is intrusive it can be likened to looking at a scene through a dirty window - when you clean it (remove the jitter) the whole picture cleans up too.

HTH - it is intended as a serious answer.

John Dawson

Re: It's a USB powered DAC implemented with asynchronous transfer protocol...

Heh - one obvious typo - ear should of course be gear - oops!

It's a USB powered DAC implemented with asynchronous transfer protocol...

...and it's aimed mostly at laptop users, so it needs low power consumption. Nevertheless you get something with around 110dB signal to noise ratio which is capable of driving headphones of any impedance to a decent sound level without audible distortion or noise. There's even a handy volume control on the box to save faffing about with a mouse. You don't get all this from any computer motherboard or laptop I know of. Sure some premium sound cards might perform as well (and quite possibly even better) but that's not what this product is competing with.

What we observed is just how many people store their music libraries on a laptop or PC, spend lots of money on headphones (I am talking hundreds of pounds) and yet don't have decent electronics to drive those headphones. To do this well for 150UKP retail is actually pretty impressive imho.

And not one of you referred to the method of ensuring proper clock recovery so that the chosen DAC chip can work to its full potential over USB. I promise you it makes a BIG difference to sound quality when sending audio over USB. It's also expensive to implement - there's a fancy XMOS processor in the box..To anyone suggesting all DACs etc sound broadly the same I suggest you read up and understand the required overall architecture before opening your mouths and going off half cock. I have given lectures and workshops to the AES on this topic.

My admission here - I was involved in the architecture of the rPAC and I am a bloody good audio electronics engineer with a deep understanding of the difficulties of analogue and mixed signal design. I also have 30Ks worth of my own ear to measure this stuff properly. On the other hand I didn't write the marketing blurb or have anything to do with the reviewing process.

Hey if you like music and have decent headphones why not try an rPAC? If you purchase it mail order you can no doubt demand your money back in 7 days if it doesn't do what it says on the tin. It won't take that long to find out ;)

John Dawson

Samsung to launch '€400' Blu-ray/HD DVD player next month

oops - rumour is wrong again (probably)

If you go back to the source of all this (in Germany IIRC), you will find it is supposed to cost 400 Euros MORE than a standard Blu-ray only player. If you think about it that makes a lot more sense and implies a price of more like 800 - 1000 Euros. Anyway we will all know at IFA next month.

See for example http://www.electronista.com/articles/07/07/04/samsung.hybrid.launch.info/

John Dawson

Defence firm loses domain case against cannabis site

remember li'l abner guys?

I thought I wasn't having a senior moment, though sadly they do get commoner these days. Way back in the 50s my auntie from Canada used to send me the Sudbury Star comic section, wherein there was a strip called Li'l Abner, featuring life in the hillbilly town of Dogpatch and including a guy who ran the Skunk Works, err... skinning skunks.

A little googling throws up the story, as featured in El Reg in 1999, called What the Hell is a Skunk Work? See http://www.theregister.co.uk/1999/10/27/what_the_hell/

Shouldn't Lockheed be suing the estate of the (presumably) late Al Capp, author of the comic strip from which they stole the name? Or should the estate be suing Lockheed? Answers on a postcard please.

All the best,

John Dawson