Looking for device upgrade for Salton ICEbox

[DPE]

I just bought the Salton ICEbox, model no. IBFS1, from overstock. It's controlled via an IR remote as well as an IR keyboard. This is last year's model, and has been discontinued. Here's the link to the new model on the manufacturer's site.

I've attempted to learn the remote with my 2117, but I cannot get anything to learn. I've tried the various combinations of short/long key presses on the learning and teaching remotes with no success. I've searched the forum here, the old yahoo forum, and the remotecentral forum as well for info.

Are there any other avenues I can explore?

Thanks,
Dave

[jon_armstrong]

Are you sure it's IR not RF. Go to a different room and see if you can change channels. If so, it's RF and can't be learned or controlled with an OFA (or any other universal) remote.

If it is IR, then try learning from the OEM remote with a sharp tap on the key. Make sure that you initiate the learning first with the OFA remote. The OFA's LED should be rapidly blinking, then a tap on the OEM remote.

[DPE]

Jon,

It is IR. It only operates within line-of-site, and I could see the IR LED blink with my digital camcorder.

I've tried learning as you suggested both with the 2117 and with a 9910. Both remotes seem to just time out during the learning process. I've tried all combinations of short/long key presses on both teaching and learning remotes with no luck.

If I could capture the pulse train with a photo diode and an o'scope, would that be helpful at all?

-Dave

[johnsfine]

If I could capture the pulse train with a photo diode and an o'scope, would that be helpful at all?

If you capture enough information at the right time scale and communicate it to us understandibly, we certainly can figure out the signal structure and almost certainly figure out how to reproduce it.

But a signal your JP1 remote can't learn is likely to be difficult to capture meaningfully, especially at a single time scale. Either the sampling rate is too slow to see the detail or the total duration is too short to see the structure. You'd probably need to capture the same signal at a few different timescales to get enough information.

We would want a timescale where the modulation (or lack of it) can be measured. Typical modulation is around 40Khz, so you could see it well with samples 80Khz or faster, but a hard to learn signal might be much faster, so I'd start at 1Mhz sampling and see what things look like there.

We would need a timescale where bit structure is clear. Typically sampling around 10Khz will show bit structure well.

We would need a timescale where frame structure is clear. A frame is probably 10 to 100 ms long. You should see whole frame(s) if you capture a 100 or 200 ms duration and then reduce to a time scale that covers just one frame.

We would need a timescale where repeat pattern is clear, that is probably a one or two second duration.

Before going to all that trouble, you may want to make sure it really is a hard to learn protocol rather than operator error in the learning process. Have you learned from more typical remotes in order to be sure you understand the process?

In case all the above time scale stuff makes no sense, o'scopes may have changed in the four years since I last used one (and obviously better ones existed for much longer than that). The best I ever used captured just 1000 samples, which seemed a silly limitation four years ago and should be absurd now, but who knows. I even had to use one limited to 128 samples. At 128 samples, it would be very hard to get a full frame duration and still have enough detail to mean anything.

[zaphod7501]

Possibly a more meaningfull result of an oscilloscope reading would be to determine if it is a learnable or reproducable signal. If the output is PWM or simple frequency generation then it is incompatible with modern remote systems. Sony's first US IR remote (and many UK remotes) was PWM and not compatible. Their PWM was basically a 60 Khz (40 or 60?) sinewave turned on and off with timings that were keyed to TV scan rates (slow). By frequency generation . I mean something like 40Khz = "on" ; 45Khz = "channel up" ; 48Khz = "volume up" ; and so on. This type of system was used with ultrasonic remote and many European IR systems for many years but was impractical for more than a few commands. Modern semiconductors may have made it possible to use this type of system again.

[johnsfine]

We can get a JP1 remote to reproduce a system based on multiple modulation frequencies. That is a good example of a system that could be read by oscilloscope and then reproduced, but could not be learned by a JP1 (or Pronto or any other that I know of) learning remote. But such systems are so rare, I don't expect that's what we'll see this time.

I don't quite follow what you mean by PWM, but there is a wide range of signal structures that a JP1 remote could reproduce but couldn't learn.

Most actual signal structures are in the much narrower range that a JP1 remote can learn as well as reproduce, so my first guess when any signal "can't be learned" is some sort of error in the learning proceedure, rather than the rarer instances of a signal the JP1 really can't learn.

[zaphod7501]

I just dug out my old remote literature and Sony called the methods burst width modulated and pulse code modulated. They are purely timing systems and contain no "data" that a microprocessor can decode. The signals were passed through bandpass filters for decoding. I've never seen any universal remote reproduce the signals but a JP1 could probably generate them. (with what I've seen lately)

The system is obsolete but a modern processor might be able to perform the filtering now , making the method usable again. I agree that it is unlikely that anyone is using this type of system but you never know what's going to show up.

[jon_armstrong]

FWIW, when I looked at the link to the ICEBox it showed a wireless keyboard and remote. My only other experience with that sort of combination is that the remote had commands very similar to the keyboard. So, if I had to guess it's learning technique coupled with the need to learn the remote with a sharp tap.

[DPE]

The functions on the remote are a subset of the functions on the keyboard. I'll try learning it again.

FWIW, I have successfully learned from many other remotes building device upgrades and filling in holes of existing upgrades. (Even worked with the JVC RX-9010 remote which I discovered sends two commands for every single-button press.)

I might go ahead with the photodiode experiment for my own curiosity anyway, as I have easy access to most of the parts/equipment here at work. All, except the photodiode, that is. Any recommendations on a part number or retailer?

[The Robman]

Just to clarify what Jon's saying. Even though the correct method of learning is to hold the teaching remote's button down for the duration, with keyboards the best method is to use short taps of the buttons instead. This is what he's suggesting that you try.