PGA2500

Heya guys... in my search for feedback on some of TI's newer pro audio devices, I was wondering if anyone tried the PGA2500 in any of their designs, wether on personal or proffesional commerical boxes?

It's very easy to believe what i hear from TI... (a bit like Ford tell their employees that Fords are the best :grin: )

I'd just like some real world feedback - good or bad.

Cheers 'n Beers

R

Never have... didn;t know about it in fact!

-I might knock up a couple of prototypes though, if there's the chance of a few samples?

Keith

fair enough --- forgive me... but I have to put on TI-Sales-mode :twisted:

You can get samples directly from TI.com - as long as your registered with my.ti

I could go into a pitch on how great my.ti is etc... but that's not what i'm here to do (this is my spare time...)

you should be able to get 3 or 4 PGA2500's through the sample program. But bear in mind, that it's like our PGA231x family -- i.e. digitally programmable gain.

You'll need something capable of spitting out a serial bitstream to control it.

Give me a shout if you want any more info -- personally, I'm looking forward to using the device in a private outofwork project. But, I'm kinda waiting for one of the guru's here to post some feedback first...

> You'll need something capable of spitting out a serial bitstream to control it.

Which is why it isn't an average DIY project.

A little like a Ford made without a battery. It just sits there until you hook it up to a jump-start. Then it runs fine all day, or until you turn it off.

Where it makes a lot of sense is in a "system" where you already have CPUs laying around, and programmers to do the ugly chores that can't be done with a soldering iron and pliers. That's generally a multi-person team, formed for-profit, and naturally not too talkative about inside details that competitors could copy.

Find out who has already bought 1,000 of them. They know what it is good for.

Agreed - I get feedback from those kind of customers... I was just wondering if anyone has done anything similar here...

Control of things on a serial bitstream is well within the DIY'ers grasp... Just look at the massive Microchip PIC community there is out there. (in fact, it's the bane of my life... :grin: )

Whether that skillset is evident with people with an Analogue Audio focus is another question. I guess that was another thing I was trying to find out in this post.

Thanks anyway...

R

I had a look at the datasheet. Programming the PGA2500 is very easy, and not a problem for any DIY'er with a bit of logic/programming experience. The small control panel I made for my relay pot circuit could easily be used for this purpose, if anybody wan't to try the PGA chip...

You can also get an evaluation board which includes programming software - but how much does that cost?

But I probably won't be testing the PGA2500 anytime soon. I don't really have a use for it - I like having a big rotary switch for gain selection...

Best regards,

Mikkel C. Simonsen

We have had a running prototype using the PGA2500 now for about 6 months using a PIC and a rotary encoder. Xavier our PIC programmer designed a small PCB and had it up and going straight away. It preforms fantastically testing wise. It obviously is using cmos or fet style internal switching which I am not a big fan of but no artefacts of this type of switching are seen when testing. It would work fine as a pure mic pre into a modern A/D chip with its +/-5v rails but to run pro +24dBM level out you need another stage running on +/-15v rails etc with a gain of 2 to get there from the +18dBM max output. The on board relay driving is a great idea and can be used for +48v, HPF, PAD & Phase switching. We Really need to do more listening tests to it to form an opinion but it is certainly is fast and clean mic pre.

Joe

JLM> fast and clean mic pre

Rochey: there's your snippet for the billboard outside the next AES convention. Joe hand-makes excellent amps, so this is high praise for a chip.

> The small control panel I made for my relay pot circuit could easily be used ... I don't really have a use for it - I like having a big rotary switch for gain selection...

There are rotary encoders, but you still need a read-out.

No reason you could not use a 12-position switch, some mickey-mouse logic, and a shift-register to walk the bits into the chip. The 12-position switch is its own read-out, non-volatile, and familiar feel. Tie the wiper to +5V, use a diode-array so each position pulls-up the desired bits, resistor pull-down (speed is not an issue), into a shift register. You could have a "Do It!" button to make the shift-register run, or maybe it can run 10 times a second if the chip does not glitch. Or get really clever with some caps to sense when bits have changed and trigger a shift.

> to run pro +24dBM level out you need another stage

Yes, but TI is betting that analog outputs and line-levels are going-away. Or that there is a market in totally-internal interfacing. For those of us who have 2 to 8 preamps and like to see wires, Pro-level will be around a long time. But TI won't get rich on us. OTOH there are those fully automated and all-digital workstations that eat mike-amps two or four dozen at a time, and sell for many kilobucks, and everything is software. The few stray analog parts need to be digi-controlled because all the controls are just software. In that world, 5V or +/-5V or p-p+/-5V is ample for an ADC sitting right next to the preamp chip. We don't have to overwhelm all the noise from here to Kalamazo (broadcast +8dBm phoneline levels) or a large recording studio (+4dBm). We do have to dodge all the digital noise inside a 95% digital box, but that is the workstation company's problem.

Think back. Data modems used RS-232 +/-12V levels, because anything less didn't overwhelm noise. RS-422(?) used +5 differential, and really does work better. And now many modems are in the same box with the CPU, interfaced as +3V just as reliable as any other signal in the PC. RS-232 +/12V is an anachronism.

There are rotary encoders, but you still need a read-out.

Click to expand...

I think digital control has its good points if used properly but can be expensive if all the extra control is only for one channel. Below is our new digital controlled Class Mic Pre which hopefully shows the good points of digital control. The control unit can be very shallow (1" deep) and controls up to 64 channels of our Class A Mic pre up to half a mile away. It uses 1 PIC and some 4000 logic to read all the encoders and run the displays.
Because of my dislike for digital in the same box as analog the PIC clock only runs while while a control is moved and goes to sleep 1second after and all displays are run in steady state and not multiplexed. All switching is done by vacuum tube reed relays. A small self powered Midi to LAN converter box can be cut into the LAN cable anywhere Protools or our own software controller can control the mic pre as it will think it is a Digipre but the 1dB normal steps can only be done in 3dB steps by Protools. Also the pre has heaps of bank memories and channels can link and track each other. Gain of -9 to 80dB in 1 dB steps etc. 90% of the features would be impossible to do without digital control

No reason you could not use a 12-position switch, some mickey-mouse logic, and a shift-register to walk the bits into the chip. The 12-position switch is its own read-out, non-volatile, and familiar feel. Tie the wiper to +5V, use a diode-array so each position pulls-up the desired bits, resistor pull-down (speed is not an issue), into a shift register. You could have a "Do It!" button to make the shift-register run, or maybe it can run 10 times a second if the chip does not glitch. Or get really clever with some caps to sense when bits have changed and trigger a shift.

Click to expand...

For the PGA2500 because our thoughts where to make a good sounding cost effective 8 channel mic pre we were actually thinking of using pots into a PIC with a built in A/D so no displays would be needed. So it would look like a normal mic pre with the pointer on the knob showing the gain but using a linear pot instead of the rev log type.

Yes, but TI is betting that analog outputs and line-levels are going-away. Or that there is a market in totally-internal interfacing. For those of us who have 2 to 8 preamps and like to see wires, Pro-level will be around a long time. But TI won't get rich on us. OTOH there are those fully automated and all-digital workstations that eat mike-amps two or four dozen at a time, and sell for many kilobucks, and everything is software. The few stray analog parts need to be digi-controlled because all the controls are just software. In that world, 5V or +/-5V or p-p+/-5V is ample for an ADC sitting right next to the preamp chip. We don't have to overwhelm all the noise from here to Kalamazo (broadcast +8dBm phoneline levels) or a large recording studio (+4dBm). We do have to dodge all the digital noise inside a 95% digital box, but that is the workstation company's problem.

Click to expand...

Yes multiple channel workstations is where the PGA2500 wins big time. But I agree Pro level wont disappear for a long time to come.

Think back. Data modems used RS-232 +/-12V levels, because anything less didn't overwhelm noise. RS-422(?) used +5 differential, and really does work better. And now many modems are in the same box with the CPU, interfaced as +3V just as reliable as any other signal in the PC. RS-232 +/12V is an anachronism.

Click to expand...

Yes we generate our own version of RS-422 so one LAN cable or spare mic cable can control the 64 channels at extreme distance with error checking all at 5v.

Joe

damn joe, thats pretty slick!

what kind of interface is on the mic pre side to accept the digital control? Im assuming you cant just go and plug any old analog pre in that box.

dave

A Small PIC PCB at the other end and eight of our new serial control Class A mic pre cards. The MP8 will be available later this year. I am working at the moment on a serial control dual 99v card and a high gain TMP8 type card which will be interchangable with the class A card. So you can have any combination of pres in the one box with the same features.

Joe

> Im assuming you cant just go and plug any old analog pre in that box.

Joe's software is so powerful, you can poke a MIDI plug in the side of an old RCA B2C preamp and his box will bend the tube grid-wires to get the exact gain you dialed-in. (uh-huh....)

No, what he has is a Good Idea that is All His. He sells a few controllers and preamps, and if people like it, he sells a LOT before his competition catches up. Depending in his intellectual property position, he can either maintain a monopoly long enough to get big market share before someone gets a "similar" but different plan going; or he can share/license the interface with the idea that if everybody is on this plan, he's bound to sell a lot of boxes even with others on the playing field.

> ....thinking of using pots into a PIC with a built in A/D so no displays would be needed. So it would look like a normal mic pre with the pointer on the knob showing the gain...

Cute, very sweet. I'm sure a well-toned coder could write the micro-program and have a dozen PICs burned and working before I could solder-up one position worth of diode-logic. And "infinite" resolution (whatever steps the chip or hand-made pre supports) instead of a 12-step 5dB resolution.

At first, I "objected" that when I want two channels "the same", pots and knobs are never as nice as a click switch. When I try to set "4.5" on two pots, I know I never quite get the pointer in the same place, the knobs are never on straight, the pots are never mounted same way, the resistance is never perfectly linear, etc. But I could strap 2 PICs off one pot, if I do that much fussy stereo. Or probably feed two pres from one PIC. So the only objection is knowing the numeric value of gain, which isn't important in recording. (A rough idea is good, but you can read a 60dB 0-10 pot to 3dB which is plenty good.) Getting LOG out of LIN is nice too, though in a sense you have just moved the problem into the gain-set matrix in the preamp. But if you want the knob away from the amp, you have to do that anyway.

[quote author="PRR"]> The small control panel I made for my relay pot circuit could easily be used ... I don't really have a use for it - I like having a big rotary switch for gain selection...

There are rotary encoders, but you still need a read-out. [/quote]
The small control panel I did does include a two digit display, and either up down buttons or an encoder. You can see it here:

No reason you could not use a 12-position switch, some mickey-mouse logic, and a shift-register to walk the bits into the chip. The 12-position switch is its own read-out, non-volatile, and familiar feel. Tie the wiper to +5V, use a diode-array so each position pulls-up the desired bits, resistor pull-down (speed is not an issue), into a shift register. You could have a "Do It!" button to make the shift-register run, or maybe it can run 10 times a second if the chip does not glitch. Or get really clever with some caps to sense when bits have changed and trigger a shift.

Click to expand...

Yes, you could do that - but what's the point? I could just use a standard chip like the SSM2019 or similar. If you don't need digital control, or 1dB steps, I don't really think the PGA2500 is the logical choice.

Best regards,

Mikkel C. Simonsen

Wow, i thought this thread was going to die off over the weekend, but I come back to all this!

Joe - that system looks SAAAAWWWWWWWWWWWWWEEEEEEEEEET! Is there a suppliers discount? :grin: :grin: :grin:

I havent' had a good look at the INA217 specs versus the PGA2500, but once again - i'm told that some of the performance of the PGA really does put the INA217 slightly in the shade.

The other sweet thing that I saw about the PGA is the GPIO pins that are on it. You can use them to witch on phantom power, invert the signal etc etc etc. Pretty well thought out.. (which is a new thing for major sc manufacturers :grin: :grin: :grin: )

Just to finish some of the other points... PRR - your right, the main market for this stuff are new age digital desks, where the PGA2500 will hook straight into the ADC (both running at 5V).

I guess if you did want a direct analogue out, you may be able to use (and I know one of you will pick a problem with this!) a OPA1632 differential opamp to add gain, and go straight down the balanced cable. Either that, or a NE5532 (to unbalanced + gain) followed by a DRV134 (unbalanced to balanced line driver).

MCS - I have a quick question for you - does the SSM2019 have a differential out? Does a differential output mean anything to you? (is there any value in it?)

Thanks again for the feedback

R

Hey! whats up with the 1632? they be in my PCM4104 dac schem atm...

Just to say tho, that there Pilo used one of those PGA chips for vol control with MIOS and a PIC, chek:

http://www.midibox.org/cgi-bin/yabb/YaBB.cgi?board=concepts;action=display;num=1081419906;start=20#20

[quote author="Rochey"]MCS - I have a quick question for you - does the SSM2019 have a differential out?[/quote]
Don't you know the answer to that one your self? The 2019 is just like the 2017, the 2015 and the INAs. Balanced input and unbalanced output.

Does a differential output mean anything to you? (is there any value in it?)

Click to expand...

I don't really care if there's a balanced output or not. A "standard" +/-15V balanced output could be useful I guess. But I don't really see why a balanced output is better than an unbalanced, it the signal goes straight into an ADC (that most likely has both balanced and unbalanced inputs).

Best regards,

Mikkel C. Simonsen

1632 is a nice little part

From the sounds of things, your using them as buffers?

As for those PGA231x chips - they really are quite good.. pin for pin with the older cirrus chips as well (and perform better). I have my own little project i'm working on that'll be using the PGA4311 (4 channel device).

I will share it once it's finished, although i don't want to quite yet... I don't want people to see how awful I am with analogue design :!:

Cheers

R

> i'm told that some of the performance of the PGA really does put the INA217 slightly in the shade..... does the SSM2019 have a differential out?

You probably should have all "similar chip" datasheets handy. Some of them are better sheets, even when they are not better chips. TI's sheet style is "sterile", uninspiring.

And these two chips reach for very different markets. The classic 1-chip mike-amps were intended to drive more analog chips, not a back-panel jack, not an ADC. It would be rare to go from MIC IN to chip directly to a rear-panel OUT.

The bulk-buyers were mixing boards, where the preamp signal went to a EQ and routing and mixing, all typically done medium-Z unbalanced inside +/-15V rails. If there were a Direct Out, on cheaper boards it was always unbalanced, and on more expensive boards it would be the designer's choice of output-jack driver.

Anyway, historically, it was a Bad Plan to put power stages on the same die as low-level stages. The large current in thin bond wires injected output garbage back into the input, the output peak heat made the input pair drift. Outputs over about 20mA 12V probably should be on a separate chip from sensitive input stages. (But they have improved this situation a lot.)

The 2500 has a very specialized output, made JUST to drive an ADC. The input of an ADC is a Sample/Hold, which may look like a high impedance but there is a LOT of glitching happening when it does its job. The stage feeding the S/H and ADC needs to have a very low output impedance far above the audio band, to suck-up those glitch spikes. Some capacitance may be part of the answer, so the driver has to swing a heavy load at high frequencies.

And I bet some of the glitching can be cancelled if the ADC has push-pull S/H (equal but opposite spikes cancel), but that requires a push-pull driver like the 2500 has. The 2500 will drive simple +5V ADCs, +/-5V ADCs, or push-pull +5V or +/-5V ADCs, whatever the system designer wants to use.

ADC designers seem loathe to work with "high voltage" (over 5V), so there isn't any call for higher output voltages.

And the low voltage allows high output current (and glitch-sucking) without excess heat or oversize power supplies.

Our only remaining concern is that it must be able to cover the input voltages, but mikes don't put out 20V peak-peak (7V RMS).

Oh, the 2500 has an awkward gap from 10dB to 0dB. So if the input can be 2V RMS (possible with some mikes), you need a pad. The on-chip relay logic is nice, avoids a major add-on design chore.

[quote author="PRR"]
Our only remaining concern is that it must be able to cover the input voltages, but mikes don't put out 20V peak-peak (7V RMS).

Oh, the 2500 has an awkward gap from 10dB to 0dB. So if the input can be 2V RMS (possible with some mikes), you need a pad. The on-chip relay logic is nice, avoids a major add-on design chore.[/quote]

PRR, youre right on the money.

The PGA2500 needs a pad for really hot mics. Although that pad can be switched out using some of the GPIO once the system realises that the mic isn't too hot.

From what I understand - the reason for low voltage ADC's is simply a case of power dissipation. The choice of silicon technology (the size of the transistors on the die etc) really drives the voltage levels that you can put through the devices.

For example - the PGA2310 (+-15V dig vol control) and it's little brother the PGA2311 (+-5V dig vol control) are actually made using two completely different silicon technologies.

The complexity involved in high performance ADC's means that there's a lot of tightly squeezed transistors etc required. To get that to a small enough size (and low power) then a smaller silicon technology is required. But, on the back of that, you have to take the lower supply voltage that comes with that technology.

I guess it's a case swings and roundabouts... you win with some features, but get a loss on another.

Cheerz

R

It is time to resurect this thread.
I have some real world info about a mic preamp based on the Texas chip (PGA2500).

I have to say, that from the first time Rockey posted this I was subjective and for me it was hard to believe this chip can produce quality results.
But last week I had access to a product based on this chip. I will not disclose the name of the product (a top company's product) but all I can say is that I was extremely surprised, in a positive way about this preamp.
It has an extremely neutral sound, very low noise and very low measurable distortion. I'm looking for a neutral solution for classical orchestral recording for years, and it looks like this chip preamp is suitable for this task.

Beeing surprised about what I saw under the hood, I contacted the designer (a well known top designer) and exchanged some emails on the PGA2500 theme. He said that he was working for some time on a discrete transparent mic preamp and then he started to experiment with this chip, and discovered that it can produce the same transparent audio quality like his complicated discrete design. In fact he said that "this is the best audio chip TI ever made".
The reason I don't want to disclose the name of the company is that I don't want to hurt his business and draw to much attention.

I thought I share this info, maybe this will gain some attention for some of you. I'm extremely interested in this project but I'm totally diletant if it is about anything related to digital/programing.

Anyone interested?


chrissugar