Finding the cmos battery

It might send its unique ID to the cloud, which remembers your favourites and tells the "radio" every time it boots?

Does it manage to remember your WiFi credentials every time you turn it on? Try a factory reset if it has one buried in the menus.

Maybe the "favourites" part of the cloud has gone away? Or is geo-restricted?

Well, if they've pulled some sort of stunt like that then it's the last time I'll ever buy a Roberts radio. It's funny they seemed to know already what the problem was since they quoted me 40 quid to fix it I'm just wondering if they don't do anything to the radio but just renew my subscription or something like that.

I've got the thing apart now and there's nothing amiss visually. There's no supercaps and no button cells or anything else of that nature. In fact the only odd thing is that there are several 220uF caps which are showing as anything between 1000uF and 1200uF on my Peak ESR70 meter. I've never known caps so far out of tolerance before. However, I don't think that could explain the issue here anyway. Sigh...

Sorry I forgot to answer your point here:

Yes, it remembers the password and accesses the router no problem.

So it has *some* functioning non-volatile storage then, wonder if it constantly writes your favourites to flash and has worn it out?

Why is this "safe to assume"?

If it has access to the internet, then it is *not safe to assume* that there is an internal power supply of some sort. It is possible that

*every* setting beyond the local WiFi AP credentials is stored on a server somewhere in the cloud. And the WiFi AP credentials could be stored in a tiny amount of flash on the main chipset to provide the bootstrap necessary to find the rest of the settings on some cloud server.

Why would you think this vs. the more typical option of the minimum wage phone worker having a table of "customer quoted symptoms" vs. "quoted repair cost" and the minimum wage worker just reading you off the quote for the item that sounded most close to your description?

In circuit or out of circuit measurement?

If they are in circult, and they are in parallel with each other in the circuit, then the measured capacitance will be the sum of the individual caps. You'd only need five parallel in-spec 220 uF caps. to show an approx. 1100 uF net value (and 1000-1200uF is "approx. 1100 uF).

The chipset /could/ contain just enough flash to store the WiFi credentials, with every other 'setting' stored in the cloud.

The bootup sequence would then be:

Boot chipset

Connect to WiFi

Download remaining cloud stored user configuration settings

Fair point. Assumptions are the mother of all fuckups - as I've said here many times myself!

You see this is all like rocket science to me. Never in a million years would that have occurred to me. If my favourites are stored in the cloud rather than locally then that's a major breach of trust Roberts has committed. And confidentiality!

Look, I'm an old dude. I have no idea what goes on in these places nowadays. I can only go by my increasingly outdated experience which is set firmly in the past (25+ years minimum).

In circuit. And they're not in parallel.

I wouldn't have thought so, Andy. It would take infintintely more write-cycles than I have time to enter data. My settings and favourites have remained largely unchanged for many years.

That is just a copout.

Realizing that you've not seen 100% of the boards in the housing is not "thinking like a technician". One does not have to "think like a technician" to realize that if you have not found what you are searching for, but have only searched one quarter of the total, that just maybe the item you seek is in the remaining three quarters.

Although your other posts have opened the very good possibility that the device has zero local storage beyond WiFi credentials.

This just requires thinking like a "cost-cutting businessman".

Question: How can we reduce the manufacture cost of this product?

Answer: Eliminate some of the components.

Question: How can we do that?

Answer: Since it already has internet access to operate, we could eliminate the local storage of settings, allowing us to eliminate a flash chip of X amount, or a battery backup of X amount.

Indeed, but none of the above means they /did not/ commit this breach for the purpose of cost cutting.

Again, think like a "cost cutting businessman". How do I reduce the cost of my "support team"? Answer: Hire the cheapest workers possible and give them a canned script to read from and follow (canned script is so I don't have to pay them for 'smartness'). The very concept of telephone support people having a "script" that they strictly follow has been the subject of jokes and parodies for a good 30+ years.

And you know this because you have the schematic or have reverse engineered the schematic from the PCB?

If you have no schematic, you are again 'assuming' they are not in parallel.

This is another unsafe assumption. It is possible that the power-down sequence could be:

detect press of soft-power button

write user settings from RAM to flash

power down

Where "write user settings" occurs whether or not you have made any changes to those settings.

And maybe the price quote you got from their phone support was really the cost for: "replace the worn out flash chip".

You can nuke its settings

press menu, main menu, system settings, factory reset, yes

also you may have linked the device to an online account via

Might your account have been locked-out?

I don't have any such account! Certainly nothing that *I* have created anyway. Unless this is something Roberts do without telling you.

... At the cost of setting up some sort of user account on the internet for the customer, which also requires time and money to implement. I don't see the saving - if any - is worth the extra bother.

Firstly, I don't believe it's feasible to 'project' (as it's called) one's own thoughts and emotions over others. You can't get inside other people's minds in this way whatever the subject matter may be. And this wasn't some "support team" lacky but rather email replies from someone who clearly understood the issue even though they were only prepared to offer a solution which would benefit the company to the tune of 40 quid.

But YOU are also *assuming* that *I* am assuming that. The fact that I am doesn't change the fact that you're still making an assumption here. It just happens to be correct on this occasion.

No one is saying *you* created anything.

All that is needed for the radio to save all your user settings in "the cloud" [1] is for each radio to have a unique serial number, and for the maker to run a very cheap server on the 'net to receive a request from the radio and to either return (for a read request) or save (for a write request) the user settings in association with that serial number.

Given that it has a WiFi chipset, it already has a unique serial number, that being the MAC address of the WiFi chip on the board.

So it is very possible for the radio to use a cloud server to store your settings. Whether it in fact does so is yet undetermined. But everything necessary on the radio hardware side to do so is already present. A WiFi chipset (for internet access), a unique serial (the WiFi chipset MAC address), and a CPU (given a WiFi chipset, it likely also contains an embedded CPU as well).

[1] yes, this is not secure, because if you learn someone's radio's MAC address, you could read/write their settings.

Ok, I had a look at the manual. Like my earlier model, it's a ridiculously complicated book more suited to a computer product than an entertainment gadget.

There is also a troubleshooting guide

In the UK, Roberts was a British brand of radio manaufacturer that catered mostly to the middle class purchaser, and I can hardly see some of those that have remaining memories of old, having some ability left to waddle through that nonsense.

However, I had a waddle. Page 64 of the wordy user manual.

"You should not attempt to update your radio unless it is recommended to you by Roberts Customer Services. Updating the software may remove all network settings, radio station presets and alarm settings from your radio"

So, station presets _are_ stored in flash memory on the device. Have they been lost in a recent software update?

Suggest, (if it still works) registering an account (Page 30) and storing favourites there.

Mr Doom. You won't find a battery or a cap. Give up.

--
Adrian C

I see. Well, I never wanted *anything* to do with 'the cloud' - ever. So if that had been the case I'd have had strong words with Roberts!

Many thanks indeed for that summary, Adrian; much appreciated. I have never - knowingly - ever done any updates at all since I bought the thing. I bought Roberts because I've had their broadcast recievers in the past and been very happy with them in all respects. They were a bit pricey (though not as much as a Hacker) but the build and sound quality was very good. Their internet radios, OTOH, have none of that reassuringly solid feel about them. I think I'll reassemble it and have a think about how to proceed from there tomorrow....

It could be possible that "Roberts" (the brand name) has gone the way of a lot of the traditional US "brand names" of yesteryear (Honeywell, Westinghouse, etc.).

They (the US brands, I have no idea re. Roberts) are nothing more than "Names" owned by some rent-seeking corporation who will sell the right to brand anything as "Honeywell" or "Westinghouse" for the right price.

What you mention re. the massive reduction in build quality is exactly what happened to the products that were "branded" as "Honeywell" or "Westinghouse" (or others) once they became a "brand name for hire". The product build quality plummeted.