IRScrutinizer INI frequency - now need help with rest of IR

J-P NYC · Post by **J-P NYC** » Sun Sep 24, 2017 6:31 pm

A few months ago I posted looking for the Pronto hex codes for a Xiaomi Mi Box (http://www.remotecentral.com/cgi-bin/mb ... ?8164,last)

With some help I was able to generate the codes using IRScrutinizer which worked nicely with my Pronto. I since have moved some components around and would now like to use my RFX-6500 (the RF transmitter that works with the Pronto) to transmit the codes, but strangely enough the codes from a few months ago dont work when they pass through the RF module. The Mi Box almost never responds to those codes when they are transmitted via RF and the front LED of the Mi Box 95% of the time doest even blink that the codes are being received.

After playing around with IRScrutinizer INI file I think the problem is that the frequency I was originally given (36k) for the MI Box codes must be off by a little bit and when it passes through the RF transmitter it must be distorted a little bit more.

I changed the INI file to use 38K and 38.4K and the MI box now consistently blinks that it is seeing an IR transmission (via the RF module), but the codes no longer function. Im sure after increasing the frequency I need to increase the pulse H & L proportionately but hvae no idea what to modify.

Can someone give me some help with this.
Thanks

Post by **The Robman** » Sun Sep 24, 2017 9:02 pm

This is tricky, without knowing what the RF box is doing to the signals, it's impossible to know which adjustments are needed. Do you have any way of recording the signals transmitted by the RF box and displaying them in Pronto format?

J-P NYC · Post by **J-P NYC** » Sun Sep 24, 2017 9:05 pm

I have 2 Prontos.
I could send it RF from one Pronto, and capture it with the other if that would help?

Barf · Post by **Barf** » Mon Sep 25, 2017 4:30 am

This is a problem with the RF transmission, that somehow "mutilates" the signal. As Rob says, the systematic way to address this problem is to analyze exactly what comes out of the RF box.

Im sure after increasing the frequency I need to increase the pulse H & L proportionately

It is unlikely that the RF links acts like a "pitch shifter", rather that it does not "like" certain frequencies. More likely is that it for example lengthens (shortens) the on-periods and shortens (lengthens) the off periods by the same amount. You can these sort of "mutilations" by subtracting (adding) an amount to all on-periods and adding (subtracting) to the off periods. For example: change

{36k,290,msb}<2,-2|2,-3|2,-4|2,-5>(1000u,-2,D:8,F:8,C:4,2,^30m)* {C=(D:4:4^D:4^F:4:4^F:4)} [D:0..255,F:0..255]

to

{36k,290,msb}<1.7,-2.3|1.7,-3.3|1.7,-4.3|1.7,-5.3>(1000u,-2.3,D:8,F:8,C:4,1.7,^30m)* {C=(D:4:4^D:4^F:4:4^F:4)} [D:0..255,F:0..255]

but it is "wild fumbling"... and I would not recommend it, except possibly as a last resort

The deployed protocol is really quite sensitive, in that it requires the decoder to tell gaps of length 4*290=1160 from gaps of length 5*290=1450; 20% difference.

Post by **The Robman** » Mon Sep 25, 2017 7:21 am

J-P NYC wrote:I have 2 Prontos.
I could send it RF from one Pronto, and capture it with the other if that would help?

Absolutely. If you do that, post a sample (in Pronto hex) of some of the buttons that you're sending, then post a sample of what they look like coming out of the RF blaster device.

Not promising we'll be able to come up with something that works, but at least we'll know what's happening.

J-P NYC · Post by **J-P NYC** » Mon Sep 25, 2017 8:18 am

I think Barf is correct, I found Robman's codes which are 38K and have the same problem (https://www.hifi-remote.com/forums/dload ... e_id=14381)

I put Rodman's code in my pronto and they also work perfectly as IR. I then tried to re-capture them from an IR emitter after they passed through the RF module and the Pronto also cant detect or capture the RF to IR. It just times out.

This also didnt work

Code: Select all

{36k,290,msb}<1.7,-2.3|1.7,-3.3|1.7,-4.3|1.7,-5.3>(1000u,-2.3,D:8,F:8,C:4,1.7,^30m)* {C=(D:4:4^D:4^F:4:4^F:4)} [D:0..255,F:0..255]

what to try next?

Barf · Post by **Barf** » Mon Sep 25, 2017 8:27 am

Do you have any possibility at all to analyze/capture/scrutinize what comes out of the RF box? Otherwise we will just blindly guess.

I posted a "blind guess" with a correction of 0.3 (this phyically means 0.3*290 microseconds). If you prefer blind guessing, you can try with corrections between, say, -1 and 1, increased in steps of 0.1.

J-P NYC · Post by **J-P NYC** » Mon Sep 25, 2017 8:41 am

How would I analyze/capture/scrutinize if the Pronto cant see it in learn mode?

can I build something quick that would read it into IRScrutinizer?

Barf · Post by **Barf** » Mon Sep 25, 2017 8:54 am

J-P NYC wrote: can I build something quick that would read it into IRScrutinizer?

You need some suitable hardware, see the documentation. But,

J-P NYC wrote:How would I analyze/capture/scrutinize if the Pronto cant see it in learn mode?

Ohhh, I do not think you said that before. IIRC, the Prontos are considered as pretty strong at learning, so likely the signal is very ... mutilated. Does the RF thingy work with other signals? Or at all?

J-P NYC · Post by **J-P NYC** » Mon Sep 25, 2017 11:16 am

the RF works with all other devises (about 10) no issues.
it is just this one set of codes it is having trouble with.

You dont see anything wacky in Xiaomi string that would cause problems?

Im trying now with blind guessing...

Barf · Post by **Barf** » Mon Sep 25, 2017 11:31 am

Please try a well known signal, e.g. NEC1 (you can enter it in Pronto as "short form Pronto", like 900A 006C 0000 0001 0C22 38C7), learn it from sending pronto and again output through the RF thingy. Post the outcome.

As I said before, the used protocol is more demanding than most. (That may be where the wackyness sits

)

Post by **The Robman** » Mon Sep 25, 2017 12:21 pm

J-P NYC wrote:How would I analyze/capture/scrutinize if the Pronto cant see it in learn mode?

If you say the IR blaster is working for your other devices, I would suggest that you try to learn some of those signals. The reason being that I suspect that you will need to experiment a little to find the sweet spot where you can learn, even the normal signals. I suspect that the IR blaster is sending out a really strong signal which is overpowering the Pronto, so you may need a distance of more than the normal 2 or 3 inches between the Pronto and the IR source.

Once you find the sweet spot where you can learn the normal signals, try learning the problematic ones.

J-P NYC · Post by **J-P NYC** » Mon Sep 25, 2017 5:23 pm

Ok did some more testing and after much playing was able to capture some data that may help.

This original NEC1
0000 006C 0000 0001 0C20 0DCF

turns into this after RF
0000 006B 0001 0000 0C41 0015

Robmans 134, 6
0000 006C 0000 000C 0027 0016 0016 002D 0016 0017 0016 0022 0016 002D 0016 0017 0016 0017 0016 0022 0016 002D 0016 002D 0016 0017 0016 01DA

comes out looking like
0000 006B 0000 000C 002B 0015 0019 002B 0019 0015 0019 0021 0019 002B 0019 0015 0019 0015 0019 0021 0019 002B 0019 002B 0019 0015 0019 01DF

Post by **The Robman** » Mon Sep 25, 2017 6:54 pm

J-P NYC wrote:Robmans 134, 6
0000 006C 0000 000C 0027 0016 0016 002D 0016 0017 0016 0022 0016 002D 0016 0017 0016 0017 0016 0022 0016 002D 0016 002D 0016 0017 0016 01DA

comes out looking like
0000 006B 0000 000C 002B 0015 0019 002B 0019 0015 0019 0021 0019 002B 0019 0015 0019 0015 0019 0021 0019 002B 0019 002B 0019 0015 0019 01DF

Well, that's pretty damn close, most devices would be ok with that. Seeing as how it's reducing the frequency byte from 006C to 006B, try starting with 006D to see if it drops to the correct 006C. The rest of the data looks close enough.

J-P NYC · Post by **J-P NYC** » Tue Sep 26, 2017 7:05 am

Just changing the frequency didnt help, but using your logic I changed the entire string with the offset and it appears to be working.

Im going to try to write a short program that applies that differential to all of the codes. Will report back.

Thanks for that GREAT idea and all of your time and effort