Transcension/Botex UP2 dimmer not retaining config

[bruce]

I have a couple of Transcension UP2 single channel dimmers. Similar to https://cpc.farnell.com/transcension/up-2-rf/light-controller-up2-single-dimmer/dp/DP29084 without the remote fader, it’s a single channel dimmer with both a fader and DMX control. 

Also available, same part number, with Botex brand.

Both have not been used since pre-lockdown.
 

One does not retain its config when the power goes off, it resets to DMX channel 1. Also revers to switch mode rather than dim. When I tested last night, the second one seemed to remebetr DMX channel, but today it has forgotten too...

 

Spec and manual (such as it is!) suggests that it should remember. Here's the Botex version, the printed transcension one is pretty much identical.

https://images.static-thomann.de/pics/atg/atgdata/document/manual/316126_c_316126_v2_en_online.pdf

 

 

Before I break out the screwdrivers, has anyone been inside these sort of units? Am I likely to find an internal battery?

[bruce]

Any thoughts anyone?

A little more experimentation. It seems that neither of these units retain their config, although the spec says they should do. I can’t remember them being like this before, but it’s been a long time since I used them.

it’s almost like I haven’t selected a “store” option in the menu. But I’m following the instructions, there are only 3 buttons, up down and menu…

had a look inside, no battery that I can see. 2 pcbs, the dimmer board on the base and a digital board at the top.

 

[DrV]

What's that 8pin chip in the top right of your board? The max485 is the DMX transceiver but the other might be an eeprom. Can you see the part number?

[Jivemaster]

Losing program after a period of disuse is classic battery failure. 

[bruce]

11 hours ago, Jivemaster said:

Losing program after a period of disuse is classic battery failure. 

That was my original thought, but there's nothing inside that even vaguely resembles a battery.

[bruce]

13 hours ago, DrV said:

What's that 8pin chip in the top right of your board? The max485 is the DMX transceiver but the other might be an eeprom. Can you see the part number?

Good spot, @DrV.

 

The writing is illegible on the chip, but I opened the second unit and it was clear.

It's an ATMLU830, which google suggests is a 16k eeprom.

Which suggests more questions - is the eprom likely to only be used for storing config values - in which case swapping it would be possible - or is it likely to have anything else on it that would prevent replacement.

And if replacement is possible, any suggestions for an equivalent part - only suppliers I can find for that particular device are bulk Chinese.

At least it's socketed...

 

 

[DrV]

I think it might well be a 24lc16. I would check that pin 4 goes to 0v, pins 5 & 6 go to pins on the microcontroller chip and pin 8 goes to +5. The other pins might have connections but where they go is less predictable. The chips are only £2.43 for 5 from RS so I'd say it's worth a punt.

As for what's stored on there, 16k bits (2k bytes) is a lot for just 2 bytes of DMX address but the dimmer has other stored parameters eg, standalone levels or sequences, switch pack mode. It may also be the same chip used on more complicated models. And the price difference between a 2kbit part and a 16kbit part is only pence. I would be surprised if there's anything in there which you have to pre-load for it to work.

One thing that occurs to me - are you exiting out of setting the DMX address (by selecting another menu option) before powering off? It might not save the address immediately as there is no Enter button to confirm.

[bruce]

Thanks Dave, I've been digging around, and had ordered up a pack of 24lc16s just before you messaged. Fingers crossed.

There's a few other parameters stored - as well as the DMX address, there's dim/switch mode, preheat settings etc, but only a handful of bytes.

And yes, I've been through all the menu options, it's not the most intuitive user interface... but after having to reset the DMX channel every night at a recent event, I'm pretty sure I've been through all the options 🙂 - at least it was on a floor followspot, and not in the air...

 

[themadhippy]

instead of a battery theres not a bloody big value cap hidden somewhere?

[DrV]

13 minutes ago, themadhippy said:

instead of a battery theres not a bloody big value cap hidden somewhere?

There's an eeprom.

[pmiller056]

Is the IC socket contact with the EEPROM slightly tarnished? Remove and re-seat the IC and check if the equipment now works correctly. It is a known problem with cheap IC sockets. Sockets like the one shown in the photo make repair easy, but they make the equipment less reliable. If you are making something and have to use an IC socket, use a 'turned pin' socket - the extra cost is worth the improved reliability, particularly if the equipment will be stored or used in a less than ideal environment.

[DrV]

Good point. Especially if it's been in storage for a while and previously worked ok.

[bruce]

No sign of corrosion or tarnish - but I removed them, cleaned with a splash of deoxit, then tested again. No difference 😞

Interestingly - when I completely removed the eeprom, the unit behaved in exactly the same way as it has done before - works normally but does not save settings. Which might suggest a faulty eeprom. 

But I'm now starting to wonder "have these ever worked properly" - I've used them several times before, but I'm now thinking "were they all one-nighters when they never lost power....."

 

[bruce]

@DrV

You were spot on. 

I replaced the ATMLU830 with a 24C16 and both units are now working. It's a bit of a coincidence that they both failed at roughly the same time, in the same way. I wonder what the external factor was - they've both been used on the same mains power, same dmx, and both were unused since pre-lockdown.

Dave, if you're ever in Glasgow and at a loose end, I owe you a beer or a coffee...

 

 

[DrV]

Frankly it's black magic how flash and eeprom remember anything at all when you look at how it's done; just a few extra electrons hanging around on the gate of a FET. Don't know why they would both fail at once though. Bad batch of chips with less than perfect hermetic seals maybe?