Multi channel attenuator withPGA2311/PGA4311

[paradigm]

It seems that I'm far from being alone in the quest for multi channel attenuation of several independent amps, so my hope is that there is someone out there who would like to help others and me to build such a device.

My problem is that I know absolutely nothing about how to design such an attenuator and know even less about programming.

That hasn't stopped me though, from trying to find the most suitable solution to my needs.

I've got a pair of four-way speakers, actually it's a five-way system, but for the fifth one, a JBL075, I will keep the passive crossover (1uF cap), since I don't think my Behringer UltraDrives will be able to improve the sound by going active on the 075's.
So, I have a four-way system, which I'm about to designate one Hypex UcD module to each driver and therefore I need at least an eight-channel fully balanced attenuator to be hooked up between the UltraDrive and the eight independent amps.
I don't need to be able to adjust the channels separately, since the Behringers I have, can take care of any level differences.

From what I've seen so far, my gut feeling tells me the Ti PGA2311 and the PGA4311, might be the best solution.
If anyone thinks I've got it wrong, please let me know.

The PGA2311 is a two-channel chip and the PGA4311 is a four-channel chip.
http://www.electronicstalk.com/news/tex/tex247.html

Please correct me if I'm wrong again, but since I need 2*4 channel attenuation, I thought it might be smart to use four PGA4311 (because I need balanced in- and outs), two on each stereo-channel, if that's possible, that is and if it's necessary at all to double the chips for a balanced/symmetrical design.
I also thought that the PGA4311 would simplify the design and PCB layout somewhat.

So, if anyone would like to make some suggestions to the design, I'm sure I wouldn't be the only one being extremely grateful for any input.

/Roland

[Franz G]

Hi Roland

you could buy a kit from Schuro:

http://www.schuro.de/preisl-vc.htm

Scroll down, you find a description in english. It is a modular system, you can combine a multichannel solution.

Regards
Franz

[Swedish Chef]

Roland,

no, you don not need four PGA4311 for 2*4 channels of differential ("balanced") attenuation. In fact, I do not think that you actually achieve a true differential interconnection that way (but I have to give this a second thought!)

The PGA2311/4311 seems like mighty nice chips like most of the BB/TI stuff. You could use two PGA4311 (or four PGA2311) paired with INA134 differential receivers and DRV134 differential drivers (also BB/TI parts) for a more or less true differential design. (see this thread Loud Speaker Crossovers and power amp placement for a discussion of "balanced" vs "pseudo-balanced") The easiest way would then probably be to use a simple microcontroller like an Atmel AVR or Microchip PIC to control the PGA:s.

/Magnus

[paradigm]

Hallo Franz!

Don't need to scroll down, since I have no problem reading German, since I'm born in Vienna!

However, I've been looking at the Schuro kits before, but never found any balanced kits.
Besides, their kit is based on the PGA2310 and I would prefer to use the PGA4311. But perhaps I somehow missed it.

Hi Magnus!

That would be great, if it isn't necessary to use four PGA4311, to obtain 2*4 channel attenuation!

I read the thread you suggested and I can appreciate your deep and profound knowledge on the subject, and what you wrote made sence to me, although I haven't a clue about the bits and pieces that makes up a fully functional attenuator and I'm sure I wouldn't be able to put them together without some help and tutoring.

I have no problem soldering and reading schematics and part-lists and putting the things together mechanically, but I couldn't tell if a design with IC's, would work or not.

You said something about samples for evaluation from Texas, would that include the PGA4311 or was that for or the DRV134 only?

Grateful for anything that would help me build one of these attenuators.

-----------
BTW! You just got to love that dish:
schkabodiboo sckaboo!
-----------

/Roland

[Swedish Chef]

Roland,

yes, TI offers free samples of the PGA4311. As most semiconductor manufacturers they do so for most of their products. (Exceptions usually include very expensive components, evaluation kits etc). You can see the free sample availability for each component under "Pricing/Packaging/Samples".

A typical all-differential attenuator signal chain with these components would be:

(1) INA134 as a differential to single-ended receiver. (Any good opamp could be used here instead).

(2) PGA4311 as the attenuator element (fed by the INA134). The attenuator is operated single-ended ("unbalanced").

(3) DRV134 as a single-ended to differential converter for driving the "balanced" cables to the amps

(4) Microcontroller for controlling the PGA:s, power supply.

For a 2*4 channel unit this would require two PGA4311, eight INA134 and eight DRV134. You could substitute the INA134:s with say two OPA4132 quad opamps and a bunch of precision resistors instead if you want. And not to be forgotten - the 16 XLR connectors at the rear!

While putting together a system like this is not very difficult it is quite time consuming. In your case it would well be worth the effort though IMHO. Keeping crosstalk, noise and other artefacts to a minimum also requires some quality time with the spectrum analyser.

Buttum leene-a, yuoor ell-JBL system deserfes zee fery best ettenooetur!
(Produced with the
ENCHEFENIZER http://www.twinpines.nl/chef/English/)



/Magnus

[paradigm]

Hej Magnus!

I went to Ti and tried to order a sample, but the "request sample" page could not be found. I'll try it later if it's only a temporary problem.

Why would I need 16 xlr's on the attenuator, when there are only eight output's?

Now it seems I'm getting closer to a feasible solution. Now it's getting real hot.

I just had to save the "chef" link.

/Roland

[nuppe]

Hi Roland

Just thought you might find these chips interesting:
WM8816

They're very similar to the PGA chips you mentioned, but with no internal opamp, so you can choose whatever amp you like.

Also there are alternatives to the ina/drv134 chips:
That 1240
That 1420

They are also releasing a new 1200 series reciever, but they're not out yet.

I'm not saying these are better, i haven't tried any of them yet, but they seem to be at least as good as the ones mentioned earlier. Just thought i'd mention them to give you more options.
I've had a hard time sourcing the Drv/Pga parts in Sweden (Elfa has the Ina134), but these are all available from Profusion in the UK.

/Andreas

[Swedish Chef]

Those "That" circuits sure look interesting. Never heard of them before.

-> Roland. Sure there are only 8 outputs. But there are 8 inputs too...
Or were you planning on fitting the attenuator inside the Behringer (is there enuf' room)?

Have you given the volume control implementation a thought? Just a knob on the attenuator, RS232-controlled by a computer or merged with a system remote control?

I did not give the TI sample page a try but most likely it is only temporarily down. Last time I ordered samples from them they processed the request within a few hours. Samples arrived by parcel service within a week.

/Magnus

[paradigm]

Thank you nuppe!
I've been there and found that there are different models available. Should I look for the 1420 DIP or the SOIC? Then there was a 1430. What's the difference?

The 1240, I didn't find, but I found some 1200, 1203 and 1206.
Which should i look for?

On the WM8816, I did not see any balanced circuit. Is that a possibility?

Would you be capable of and willing to suggest a good design around the WM8816?


-Magnus
I still can't go to the PGA4311 "Request sample" page. Do you know if there is another way?

/Roland

[Swedish Chef]

Roland,

the 1200, 1203 and 1206 differ in their gain. (1200 = 0 dB, 1203 = -3 dB, 1206 = -6 dB). They all look great because they have a special topology that boosts common-mode impedance to closely mimic a transformer. The 1240 series is the standard solution (one opamp used in the "instrument amp" configuration). They also come with different gain settings.

Regarding the WM8816, no there is no balanced circuitry here. Neither is there for the PGA4311! They both operate in single-ended (unbalanced) mode. This is what you use the INA:s and DRV:s for. First you convert the differential (balanced) signal to single-ended (unbalanced), then the attenuator performs the necessary attenuation, then you convert the signal back to differential again.

So all in all, the WM8816 and the PGA4311 is pretty much the same thing. Basically, the only difference is that the WM8816 does not have any internal opamps necessary for proper operation. The PGA4311 also seems to have superior specs and the "soft step" function.

Bottom line is that you can use any combination of the above chips or even something else. The truth been told, I have not looked into digitally controlled attenuators for audio to any large degree. But the PGA4311 has such good specs that it would be pretty hard to beat with any discrete solution.

Now, PGA4311 sample problem... Go to the bottom of the tech page for the PGA4311 and click "Request samples". Now you have to log in as a TI user. Assuming that you are not yet registered at TI you have to fill in a registration form (shipping address, name and so on). Then there should be no problems requesting samples.
-Oh, assuming of course that you did not select "Afghanistan" and "Missile Guidance" as for the intended application.

/Magnus