Universal buffer/headamp based on OPA1622

[OlegSh]

Thanks John and ammel68!

I'll put the OPA1642 back in the cart I also updated its datasheet in my collection. Are there other op-amps recommended for the buffer after the volume pot? I have around 20 Euro to spend before I reach 50 Euro minimum order for free shipment at mouser, so I can use them for more op-amps

 

[Joseph K]

OPA827? But it's single.

Ciao, George

 

[OlegSh]

Thanks Joseph! I have single op-amp PCB design as well, I just did not manufacture it yet, but I can give it a try. Is there similar dual op-amp alternative?

 

[johnc124]

There is not a dual of the OPA827 unfortunately.

Some other op amps I'll through into the mix are OPA1602, OPA1652, and OPA1688 if you're looking to try a few variants.

 

[OlegSh]

Thanks John! I'll add a couple of each to the cart. I have enough PCBs to try them

 

[ammel68]

I wish someone could put some '1688s on DIP-8 adapter boards so more people, including myself, could try it.

 

[Joseph K]

A very fine sounding FET input opamp is the ADA4627, really in the same leauge like the OPA827, but different kind of sound.
Then I don't know why, but the ADA4610, ADA4001 do not get so much publicity in audio circles. They look very promising, too? No personal experience, too, but they make me curious? Just thinking loud..

 

[OlegSh]

ammel68, I understand the trouble with the SOIC packages when it comes to hand soldering but SOIC to DIP adapters have such long traces that I prefer not to use them. All my PCBs are designed for SMD parts since most of new interesting op-amps are in SOIC or even smaller packages. And once you allow yourself one SMD part then why not most of them since it makes the layout really tight. If you can find someone near by who could help you soldering SMD parts I can send you one of my dual op-amp PCBs. I've implemented extra pads for the compensation scheme (series RC network from -IN to +IN) proposed by John (johnc124) for OPA1688, but it needs 0603 parts, the other parts are 0805 and a couple of 1206 output resistors.

Joseph, thanks for interesting parts! My 20 Euro to spare are almost gone now but I'll keep analog op-amps in mind for later experiments.

Regards,
Oleg

 

[OlegSh]

I’ve just warmed up my soldering iron to solder the OPA1642 to the buffer PCB but decided to check its datasheet once again. I found out that OPA1642 has 1mV typical (3.5 mV max) input offset voltage (how did I overlook it before!?). Combined with the gain of two of the following OPA1622 circuit it may result in 2~7 mV DC at the output. I understand that even 7mV DC at the output would result in around 1.5 uW of power into 32 load so it won’t fry the headphones. But I’m still in doubt regarding other negative effects that I may have overlooked. So should I worry about such a output offset voltage or it really makes no harm?

Regards,
Oleg

 

[agdr]

Suggestion: use an OPA2140 instead of the 1642. Offset gone.

 

[OlegSh]

Thanks agdr!

I just checked the specs of the OPA1642 and OPA2140. Surprisingly the specs are nearly identical except for the input offset voltage and current. Even noise and THD+N plots vs frequency are nearly identical. OPA2140 has also lower overshoot with the same output isolation resistor. So OPA2140 it is Time to place another order at mouser…

 

[agdr]

The OPA2140 is a great chip. Credit where credit is due, it was Sergey888 who first made me aware of the OPA140 several years ago when I came up with the O2 Booster Board. At that time the 2140 and 4140 (dual & quad of the 140) were not shipping yet.

The only specification about the 140/2140/4140 I've had to ponder a bit over the years is the open loop output impedance vs. frequency, figure 18 in the datasheet. That shows output Z shooting up to 1K as the frequency drops from 1kHz to 10Hz. Compare that to the same graph for the OPA827, figure 27 in that datasheet, nice and flat 20 ohms out to10MHz.

http://www.ti.com/lit/pdf/sbos498 (OPA140 datasheet, opens PDF)

http://www.ti.com/lit/pdf/sbos376 (OPA827 datasheet, opens PDF)

My best understanding of it all - this shouldn't matter for the application where the 2140 is looped around another chip, such that the load on the 2140 is nearly zero. 1uA and whatever the OPA1622 input capacitance is in this case. The outer loop negative feedback should take care of the output Z. Or at least it would if it were a single chip with negative feedback. Where I have to ponder things a bit is what happens to that 2140 output Z vs.frequenc with the negative feedback around the looped pair, OPA2140 + OPA1622?

I don't really have an answer to that question, which is why I've stopped recommending the OPA140 in my Booster Board project (OPA827 + LME49600) in favor of the OPA827. Maybe johnc124, Sergey888, or anyone else who is a lot more current than me in feedback/control systems like this can chip in.

 

[ammel68]

Never heard of the OPA2140 before. I'll have to get one to try.

 

[chris719]

LME49720 is a fine op amp and I don't think you'll be disappointed with it.

With regards to the OPA1642 distortion with higher source impedances, it's actually one of the best performing op amps I've ever measured in that regard and will outperform the LME49720 with high source impedances.

The old datasheet for the OPA1642 didn't show how good the performance was, and an improvement was made to the process after the datasheet was released. The new datasheet reflects how low distortion will stay even with a 10kOhm source impedance

See figure 37: http://www.ti.com/lit/ds/symlink/opa1642.pdf

Either op amp is a great option, enjoy your project!

Regarding Figure 37 and the process tweak; does the same apply to the OPA827?

 

[OlegSh]

The only specification about the 140/2140/4140 I've had to ponder a bit over the years is the open loop output impedance vs. frequency, figure 18 in the datasheet. That shows output Z shooting up to 1K as the frequency drops from 1kHz to 10Hz.

After you suggested the OPA2140 I compared its open loop output impedance with that of the OPA1642 and they behave similarly except that the later has deeper minimum between 1kHz and 1MHz. I guess in my case (unity gain buffer before the OPA1622) it should not matter much since the closed loop output impedance of the OPA2140 should be fine.

 

[OlegSh]

Got the parts from mouser! Now I've got one more question before I turn on my soldering iron Is input impedance matching as important for the JFET input op-amps as it is for BJT input op-amps? How critical it is for OPA1642 or OPA2140 in unity gain configuration? So far I plan to just jumper the output and inverting input of the OPA2140 for lower noise.

Thanks,
Oleg

 

[johnc124]

Regarding Figure 37 and the process tweak; does the same apply to the OPA827?

Yes, OPA827 is on the same process as OPA2140 / OPA1642 and will show essentially no additional distortion with a high source impedance and in a unity gain buffer configuration.

 

[johnc124]

The OPA2140 is a great chip. Credit where credit is due, it was Sergey888 who first made me aware of the OPA140 several years ago when I came up with the O2 Booster Board. At that time the 2140 and 4140 (dual & quad of the 140) were not shipping yet.

The only specification about the 140/2140/4140 I've had to ponder a bit over the years is the open loop output impedance vs. frequency, figure 18 in the datasheet. That shows output Z shooting up to 1K as the frequency drops from 1kHz to 10Hz. Compare that to the same graph for the OPA827, figure 27 in that datasheet, nice and flat 20 ohms out to10MHz.

http://www.ti.com/lit/pdf/sbos498 (OPA140 datasheet, opens PDF)

http://www.ti.com/lit/pdf/sbos376 (OPA827 datasheet, opens PDF)

My best understanding of it all - this shouldn't matter for the application where the 2140 is looped around another chip, such that the load on the 2140 is nearly zero. 1uA and whatever the OPA1622 input capacitance is in this case. The outer loop negative feedback should take care of the output Z. Or at least it would if it were a single chip with negative feedback. Where I have to ponder things a bit is what happens to that 2140 output Z vs.frequenc with the negative feedback around the looped pair, OPA2140 + OPA1622?

I don't really have an answer to that question, which is why I've stopped recommending the OPA140 in my Booster Board project (OPA827 + LME49600) in favor of the OPA827. Maybe johnc124, Sergey888, or anyone else who is a lot more current than me in feedback/control systems like this can chip in.

OPA827 is a classic emitter-follower output stage, while the OPAx140 / OPA164x devices are rail-to-rail outputs which tend to have higher open-loop output impedances. The behavior at low frequency is also shaped by a miller loop closing around the output stage. However, it's important to remember that the closed-loop output impedance is the open-loop output impedance divided by (open loop gain * feedback factor) + 1. So closed-loop output impedance is almost always extremely low at audio frequencies.

 

[OlegSh]

Hi John,

I hope you are still around so I can catch your attention I compared fig. 9 and fig.37 in the OPA1642 datasheet and the THD+N with zero source impedance seems an order of magnitude lower at 5Vrms input in unity gain configuration then it is with 10k source impedance. Am I right to assume that while there is no frequency dependent distortion rise due to constant input capacitance, the impedance mismatch at the op-amp inputs would still result in increased distortion or I’m interpreting the figures incorrectly?

Thanks,
Oleg

 

[johnc124]

There is no increased distortion. With regards to Figure 37, that was measured with the amplifier in a buffer configuration (not the distortion gain configuration shown in Figure 35) which means that the noise floor of the audio analyzer (APx555) limits the measurement capability. But also remember the "N" portion of THD+N, you will always have an increased noise due to the source resistor itself. In this case however, the noise floor of the audio analyzer is the major contributing factor.