Si4711 DSP Modulator with Stereo & RDS

[Maximus]

Hi necks.
I stumbled upon these ics a while ago and thought to have a look and try to see how they work. They look easy enough to implement.

Has anyone tried these before as I have a feeling these might be what's inside the radiomaster rigs?

There's some clever people out there that seem to have had quite a bit of success with them.

More info:
http://www.silabs.com/products/audio/fm ... 71011.aspx

http://tipok.org.ua/node/35

http://vrtp.ru/index.php?showtopic=21286&st=0

http://www.elv.de/hq-stereo-ukw-pruefge ... zteil.html

There's also some demo videos on YouTube and fancy building one myself when I get a bit of time.

Was just wondering what everyone else's views are before I invest any time.


Sent from my iPhone using Tapatalk

[OgreVorbis]

All Si47xx have the same shitty performance. They create spurs everywhere. Better off with a BH chip. They do sound nice though and have a lot of features. They are a pain in the ass to solder though. I've tested them personally on a spectrum analyzer. The spurs are -30dBc at worst, but there is a "second carrier" -20dBc at about -200kHz from carrier. Not worth it. The newest PCS STMAX transmitter uses these, but all his other transmitters use high quality independent VCO and PLL.

Radiomaster is far ahead of this. He uses a separate PLL MC145170 and some really advanced VCO he designed himself.

There are a few all-in-one ICs that I still haven't tested: SN761634, BK1085 (easy to solder), and QN8007. After those three are tested, that's all the FM chips on the market - I researched hard. If someone could help me test some of these and report the spectrum, that would be very helpful because it's hard for me to obtain, solder and program all these chips just to find the spectral purity.

The one I am most interested in at this point is the QN8007 because it is the only one other than the Si chip, that reports its spectral purity in the datasheet and it says -45dBc at worst point (which is very good for an all-in-one chip).

Here is a quote of my post from the other forum:

In order from best to worst:
1. BH1414K: >-60dBc across band, -50dBc +-200khz (decent sound quality, least spurious)
2. RDA5820: -45dBc across band, -60dBc +-200Khz (best sound quality, but ONLY WORKS FOR 5 MINS, THEN PLL RANDOMLY UNLOCKS)
3. BH1415F: -50dBc across band, -20dBc +-200Khz (but sound quality is worst)
4. Si4713: -30dBc, -25dBc +-200Khz - has a double carrier
5. AS6001/2: -25dBc (fills most, but not all of band) - also has a triple carrier -200 and -400khz
6. KT0803L/K: -25dBc (fills entire FM band with spurs)
7. QN8027: mostly -25dBc some -20dBc

I'm pretty sure that's all the chips on the market. Let me know of any others.

RDA5820 would probably be the winner if it didn't keep unlocking. Audio quality seems best on this one as well. Sounds exactly the same as the source signal to me, but in terms of spurious radiation the BH1414K is best.

Read more: http://darkliferadio.proboards.com/thre ... z4MARKf08z

[OgreVorbis]

See datasheet for QN8007: http://www.datasheetbay.com/pdf/1028100/QN8007B.html

Compare to Si chip: http://www.silabs.com/Support%20Documen ... 11-B30.pdf

See "transmit channel edge power" for Si and compare "RF output spectrum mask" for QN. It also has an integrated band pass filter which no other chips have. It looks very promising. I also began designing a test board for it. It is probably the only one suitable as a real replacement for the classic PLL + VCO transmitters if it lives up to the specs.

[yellowbeard]

I think the one chip transmitter suffers from an obvious flaw, and its one we all probably encountered - keeping digital noise like clock signals and their derivatives from being modulated and amplified in the RF part of the transmitter. Screening and careful design are how you deal with it old school but when its all on a single chip you mostly only have careful design. It would be great if you could bypass the oscillator in some of these chips and get the modulating signal for stereo/RDS on your proper, clean man size transmitter.

[Albert H]

I played with a few of these ICs, and they all were pretty disappointing. In every case, no matter how they were carefully decoupled, the digits would always turn up in the carrier. They're fine for their intended purpose - connecting your iPod to your car radio - but they're no good for broadcast use (at least the few I've tried).

I do use an integrated VCO on one driver board I build. It was a convenient shortcut that allowed me to hide the VCO inside a little screening box (that I accidentally bought a couple of hundred of.....)!

To overcome the digital filtering issue, I have a double-sided PCB, with the PLL logic on one side, and the RF bits on the other, with a complete groundplane. The carrier power is 5 Watts (variable) and the worst spurious output is -90dBc - there's just nothing where it shouldn't be! Electronically, it's a simple broadband design, but the real secret lies in the layout. The latest PCB is three layers on a double-sided board, and is 45% smaller than the original!

[OgreVorbis]

I played with a few of these ICs, and they all were pretty disappointing. In every case, no matter how they were carefully decoupled, the digits would always turn up in the carrier. They're fine for their intended purpose - connecting your iPod to your car radio - but they're no good for broadcast use (at least the few I've tried).

I do use an integrated VCO on one driver board I build. It was a convenient shortcut that allowed me to hide the VCO inside a little screening box (that I accidentally bought a couple of hundred of.....)!

To overcome the digital filtering issue, I have a double-sided PCB, with the PLL logic on one side, and the RF bits on the other, with a complete groundplane. The carrier power is 5 Watts (variable) and the worst spurious output is -90dBc - there's just nothing where it shouldn't be! Electronically, it's a simple broadband design, but the real secret lies in the layout. The latest PCB is three layers on a double-sided board, and is 45% smaller than the original!

The "integrated VCO" - is that the one from minicircuits?
It sounds interesting. I'd like to see a pic of your board.

[Albert H]

My board isn't too interesting. It uses a length of RG176 coax for the oscillator resonator, a PLL IC with an active loop filter, a PIC, a FET for the oscillator, another FET for the buffer, then a couple of bipolar transistors with lots of negative feedback, driving a Mitsubishi 6 Watt FET configured to be broadband. There's an RF level detector (a diode pump) that is used to tweak the drive, so that the power is flat at any frequency, since some of the PA devices I've used lately are completely intolerant of over-drive. There's also copious screening, and over-temperature detection (the crystal reference can go a bit mad at elevated temperatures). The varicap circuit is scaled so that peak mod (75kHz deviation) equates to exactly 1.5V p-p at any frequency, so no adjustments are necessary. Frequency selection is by DIP switch. The latest version of the board is 10cm × 5.5cm and fits exactly into a standard tin plated screening box.

[Analyser]

The "integrated VCO" - is that the one from minicircuits?
It sounds interesting. I'd like to see a pic of your board.

I don't think Albert has a camera.

[RF-Head]

Maybe your wright i never seen any picture from Alberts projects
He is only talking about many big projects he made in the past
So please Albert show us some nice picture of your projects

[Maximus]

Thanks for the replies gents. I kind of thought as much as with regards to the sprogs and digital noise. Seen it many times on the BH ICs. That's why I've never used one properly apart from testing


Sent from my iPhone using Tapatalk

[Albert H]

Here's tonight's project - a quick 10W carrier, 50W peak MW box. It's needed in France on Sunday, so I'd better get cracking. I got the synthesiser built in about an hour. The rest should take another two hours or so.

[Analyser]

Here's tonight's project - a quick 10W carrier, 50W peak MW box. It's needed in France on Sunday, so I'd better get cracking. I got the synthesiser built in about an hour. The rest should take another two hours or so.



(Just right-click on "Image" above, and select "view image"). This part of the site doesn't seem to allow in-line images.

Excellent, looking forward to seeing some VHF gear soon.

Using an image hosting site such as imgur seems to be the way to go so that pictures appear in posts.

[SD-E1102]

It does work with that site using the correct direct link, all fixed.

[Albert H]

Incidentally - have you noticed the "gimmick" capacitor adjusting the crystal? It's the twisted white and blue wire. You twist a length of insulated single strand wire, then cut it so that you just have two bits of wire twisted together. The ends are connected into the circuit, and the capacitance it provides is reduced by shortening it. It's a lot cheaper and much more accurate than a trimmer capacitor! The bit that's used there is about 2.547pF.

[Albert H]

Excellent, looking forward to seeing some VHF gear soon.

The VHF and UHF stuff these days is mostly "box-shifting" - buying in redundant plant from Yank stations, giving it the once-over to make sure it's working properly, converting to 230V AC if needed and then selling to customers in Eastern Europe, Africa and Asia. It's quite a good business, but you do need a couple of highly competent engineers who don't mind a lot of travel to do the installation and commissioning of the gear.

I do have a couple of UHF things on the bench at the moment (link gear), and I'll do a "show 'n' tell" when I have the time to take some photos.