My TAS5630 build - questions inside.

[Jmz1]

I got a couple of free samples of TI TAS5630PHD chips, as well as 4 OPA1632 precision differential op-amps.

I have a scavenged Avalon C150 preamp minus its transformer, which I plan to use as a case, as well as scavenging its RCA connectors, power switch, and external pot knobs.

I was thinking about using something like this for a passive preamp:

* 2 pcs * of DACT Type 21 Stepped Attenuator 100K 2A3 - eBay (item 200451943693 end time Apr-18-10 06:52:20 PDT)

Would I be better off without a preamp? I will be using this as a portable amplifier, to be plugged into iPods and such. I figured an attenuator would give me more flexibility in the line levels and noise susceptibility.

I also have the option of going for an active preamp, using the OPA1632s.

My plan to power it is initially to use an external 50 V DC supply, then designing a DC-DC converter to provide 50 V from 12 V and allow use from a car battery. Hopefully this can also be integrated inside the preamp case, as it is quite large. I would be concerned about EMI from the converter though, so there would possibly have to be some shielding.

I have noticed a conspicuous absence of ESD protection in all the chip amp designs I have seen, is this considered ok? You've got an externally accessible input going fairly directly to the pins of an IC, I would have thought there would be some issues there.

I'll have to track down some inductors, Ice Components are looking good.

Thanks for reading, any pointers are appreciated.

 

[Jmz1]

Good news, I found a bunch of (free) 680 uF 80 V capacitors in a nice form factor for the power supply decoupling, I will use two for each half-bridge power input. I was thinking of also adding a 2 uF tantalum or film cap even closer to each half-bridge's power input for higher frequency decoupling, is this useful? The reference design doesn't use anything of the sort, just a single 1000 uF.

I also have samples of the Ice Components 1D17A-100M coming my way. I will try and mount them properly, though the users of them I have seen here tend to have them sticking up on wires, is there any reason for that?

My preamp designing is coming slowly, I saw a possible candidate design in something like this:

http://homepage.mac.com/tlinespeakers/hhaller-passivepre.gif

from here:

Passive Pre - AudioKarma.org Home Audio Stereo Discussion Forums

The TAS5630 input resistance is only 33K. I might consider going for a lower impedance attenuator section. There is also the question of the interaction with the AC-coupling 10 uF input caps, and the 100R/100pF low-pass filter section in the reference design.

Does anybody have any recommendations for input AC-coupling caps, or the attenuator section in general?

 

[bob91343]

I am also getting a couple TAS5630 chips. That thing is amazing, 600 Watts of audio with such a high efficiency! And the price is only about six bucks.

Let's compare notes on this; I don't know how far I'll get with my project but I have some ideas of what to do. My first job is to get the thing working in the first place. I would like some of those big electrolytic capacitors to build the 50V supply.

It also requires 12V for the low level.

 

[81bas]

I was thinking of also adding a 2 uF tantalum or film cap even closer to each half-bridge's power input for higher frequency decoupling, is this useful? The reference design doesn't use anything of the sort, just a single 1000 uF.

The reference design says, that these ceramic 2.2uF caps are required to be placed so close as possible to the TAS5630 pins. Otherwise the amp will have parasitic oscillations and will produce insane noise. (I have experienced it already ) Solid ground plane on one of the PCB sides 'must have' also.

Also, if you going to use it in BTL or PBTL mode, then OPA1632 are required, because a so called 'differential' inputs by TAS5630 are done so, that each input wire simply goes to it's own hafbridge. So you cannot connect one of the TAS5630 'differential' input wires to the ground, and other to attenuator. This will produce wrong result. For this purposes OPA1632 must be used before TAS5630 inputs. Input decoupling caps and 100R/100pF filter are required too.

 

[Jmz1]

The reference design says, that these ceramic 2.2uF caps are required to be placed so close as possible to the TAS5630 pins. Otherwise the amp will have parasitic oscillations and will produce insane noise. (I have experienced it already ) Solid ground plane on one of the PCB sides 'must have' also.

Also, if you going to use it in BTL or PBTL mode, then OPA1632 are required, because a so called 'differential' inputs by TAS5630 are done so, that each input wire simply goes to it's own hafbridge. So you cannot connect one of the TAS5630 'differential' input wires to the ground, and other to attenuator. This will produce wrong result. For this purposes OPA1632 must be used before TAS5630 inputs. Input decoupling caps and 100R/100pF filter are required too.

Thanks for that, I had managed to confuse myself into thinking that the 2.2 uF cap wasn't in the reference design because I was contemplating putting an capacitor with an even higher self-resonant frequency in there.

I was also confused about the half-bridge inputs, I thought that the AC-coupling would make that OK. Evidently not.

Am I correct in thinking that I should use an OPA1632 with the common-mode voltage reference connected to ground, decoupled at the output of the OP1632, but with the differential input to the OPA1632 DC coupled?

Getting the OPA1632 supplies is going to be interesting, mostly because of their power draw. The quiescent current at 15 V is quoted at 17 mA per channel (per each +/- supply?), which is 500 mW.

My preference would be to regulate this off the 50 V bus, which means dumping 600 mW somewhere per channel. There is also the question of the negative supply, my current idea is to first regulate down to some voltage a few volts above my desired negative rail voltage (haven't chosen between 12/15 yet), then use a 20V-capable inverting charge pump, then regulate again to the negative rail voltage.

 

[81bas]

Am I correct in thinking that I should use an OPA1632 with the common-mode voltage reference connected to ground, decoupled at the output of the OP1632, but with the differential input to the OPA1632 DC coupled?

Simply do the connections like shown here: http://focus.ti.com/lit/ug/slau287/slau287.pdf (page 21-22)

 

[m-tech]

Also, if you going to use it in BTL or PBTL mode, then OPA1632 are required, because a so called 'differential' inputs by TAS5630 are done so, that each input wire simply goes to it's own hafbridge. So you cannot connect one of the TAS5630 'differential' input wires to the ground, and other to attenuator. This will produce wrong result. For this purposes OPA1632 must be used before TAS5630 inputs. Input decoupling caps and 100R/100pF filter are required too.

Are You sure about this ? Because I made bridged amp with TPA3122. Input signal goes only to first amp; input pin of second amp is grounded trough capacitor ( it gets signal trough feedback from first amp output ). So why this shouldn't work on TAS5630 ? ( You might be wright but I'm just asking ).

 

[theAnonymous1]

Are You sure about this ? Because I made bridged amp with TPA3122. Input signal goes only to first amp; input pin of second amp is grounded trough capacitor ( it gets signal trough feedback from first amp output ). So why this shouldn't work on TAS5630 ? ( You might be wright but I'm just asking ).

The half of the bridge that is AC coupled to ground is just sitting idle while the other half is doing all the work.

If you take your TPA3122 amp and AC couple a speaker from each half-bridge output to supply ground, you will notice it only play from one output.

 

[Jmz1]

Simply do the connections like shown here: http://focus.ti.com/lit/ug/slau287/slau287.pdf (page 21-22)

Ah, a single sided supply for the op amp. The common-mode reference is left floating I see, which the datasheet says means that it defaults to mid-supply.

What is the significance of the J45/J25 jumpers on the input on page 22? I see that you should connect ground to input low for a single ended input, but I don't see what should be done for a balanced input.

 

[81bas]

Input signal goes only to first amp; input pin of second amp is grounded trough capacitor ( it gets signal trough feedback from first amp output ).

You do not have any access to TAS5630 feedback path, each channel has it's own feedback path in this IC. So it is needed to feed it's bridge inputs with truly differential voltage using OPA1632

 

[81bas]

What is the significance of the J45/J25 jumpers on the input on page 22? I see that you should connect ground to input low for a single ended input, but I don't see what should be done for a balanced input.

for balanced inputs you need to remove J45/J24 jumpers and connect cable directly to the pins (instead of 'phono socket').
Also, why do you need to use balanced inputs? Do you have any device with balanced outputs? Do you plan to use any pot/attenuator at inputs then?

 

[Jmz1]

No, just hypothetically speaking.

 

[m-tech]

You do not have any access to TAS5630 feedback path, each channel has it's own feedback path in this IC.

OK, but TPA3122 also doesn't have access to feedback path, and everything works fine. But this probably wouldn't work with TAS5630 ( maybe different architecture compared to TPA3122 ? ).

 

[81bas]

OK, but TPA3122 also doesn't have access to feedback path, and everything works fine.

No, it can work better If you will give inverted audio signal to the second input (instead of grounding it), then you will get 2x more power undistorted... Look into the post #8 in this thread

 

[theAnonymous1]

OK, but TPA3122 also doesn't have access to feedback path, and everything works fine. But this probably wouldn't work with TAS5630 ( maybe different architecture compared to TPA3122 ? ).

I used a DRV134 to bridge my TPA3122D2 boards......

http://www.diyaudio.com/forums/class-d/124318-my-tpa3122d2n-btl-proto-6.html#post1872915

The DRV134 was configured for a single supply and I don't think it likes it; there is some unwanted noise present.

I'm going to redo the layout once again to use the OPA1632 instead. I will be using the same OPA1632 circuit that is in the TAS5630 eval board datasheet.