Op-amp based buffer amplifier

[Reffc]

I've been meaning for a while now to build a buffer amplifier which I can place between any preamp and any power amp to ensure no more impedance mis-match issues. My current power amps have a low-ish input impedance of 5K1 Ohms which should still be fine for the croft pre (300 ohm output impedance) but I think that by using a buffer amp, it'll make the Croft's job easier.


The easiest way (so I was advised) to achieve this might be through the use of an op-amp based circuit using a unity gain op-amp, so I chose the TexasInstruments low noise dual NE5532 op amp chip for the circuit.


This needs DC power, so instead of going to the expense (for now) of a 240v to 9vDC power supply, I decided to power it with two 9V batteries.

Down-stream of the inputs, I need to ensure sufficient shunt impedance between signal and earth, so a 1mOhm resistor was picked. This was followed by a 1uF cap in series, a 1K resistor (in series) then the input bias resistor to set input impedance.

As it's the croft, 100K seems far more valve amp friendly than the 5k1 of the Albarrys, so that's what I've used. just prior to this as a precaution, there'll be shunt diodes to ensure that if slight over-voltage (DC) makes it onto the signal, it'll be dumped to the voltage rails and passed through the internal resistance of the batteries where hopefully no damage can be done.

That's the plan....here's the execution:

Bits all arrived this morning:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7315_zps0390c22a.jpg


Neat little extruded case which I figured would be just the right size although batteries will have to be sited externally and a power lead run through the front panel;


Simple 2-pole rocker switch for power:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7318_zpsdd5f0d21.jpg

Being wired up:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7317_zps3afdf401.jpg

front and rear panels drilled and RCA sockets fitted. Front panel cut out and drilled for rocker switch and power-on LED; Switch and LED fitted:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7319_zps926b44a0.jpg

Rear view:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7320_zps70ff7397.jpg


LED being tested with batteries connected:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7322_zpse6557f12.jpg


Rear view: resistor is there to prevent too much power reaching LED as it only requires a tiny voltage/current. I've used an 8K2 value:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7323_zps02bb39ea.jpg


Starting to populate the board. Left channel being figured out here:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7325_zpsc62a579e.jpg


http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7327_zpse794174b.jpg


Both channels now populated after a few hours work. Top view:


http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7329_zps27716438.jpg



Underside showing solder connections and star-earth bare copper wire roughly central to the board. Anyone spot the two mistakes? :scratch: (needs careful examination of top view!)


http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7331_zps976ffe33.jpg


I had shorted one of the series caps :doh: and hadn't connected one of the shunt resistor legs to the star earth...testing brought this up later.


http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7332_zps30d309ae.jpg


Almost completed, minus the top, and ready for testing:


http://i1140.photobucket.com/albums/n577/24pacman/Hifi/_MG_7333_zpsfe008171.jpg


Test measurements:

input impedance across signal to earth = 1MOhm (both channels)


DC offset voltage at output = 93mV one channel and 94mV t'uther channel.

This is a little higher than expected but is just the way the chip performs...a little DC is being seen relative to ground but it's still tiny and equal in both channels, so I'm happy. Another unity gain chip may have better performance but this one's well respected as a decent low noise chipset which wont add anything like significant signal distortion (perhaps not even measurable) so it's a good solution.

As gain is 1:1, all this amp does is to allow the plugging in of ANY preamp, even a passive or a standard gain pre with perhaps high output impedance, safe in the knowledge that it will drive any power amp with ease, and drive long interconnects without fear of rolling off frequency extremes. In theory it should restore an ideal balance...it's a sort of match-maker:lol:

Time to test it out :stalks:

[Reffc]

Just tried it out to start with between CDP and pre....yuck...distortion! Whipped the top off the buffer amp and ...well lets just say I'm a numpty :o I'd forgottton to star-earth the central battery connections (wires connecting positive of one to negative of t'uther battery). These needed grounding to provide 9V across switch. Connected a chassis earth while I was at it.

DC offset now a shade under 15mV which is more like it.


Now plugged between pre and power amp and wow, what a change! More bass for a start, nicer rounded sound with any hint of upper lean-ness removed. lovely quiet operation too, no noise, no hum, very very silent noise floor. Very pleased with this little box of tricks.

If any one else is contemplating a passive preamp build or wants to buffer their valve pre into whatever power amp, this little design will do the trick very nicely.

[Puffin]

I made a Pedja Rogic buffer some years ago using the same boards that are shown in the link below. At the time I preferred my X-10D valve buffer. In what way does this differ from your NE5532 buffer? I must dig it out ( I think I have some more boards somewhere as well).

http://www.justblair.co.uk/Amplifier/building-a-pedja-rogic-buffer.html

[Reffc]

Hi Rob

It's a very different circuit in terms of values. The circuit I used is detailed in the sketch below to detail the components and values...it's just doubled for both channels with a single star earth but retains just the two voltage rails:

http://i1140.photobucket.com/albums/n577/24pacman/Hifi/paulbuffer_zps4a329439.png


The only thing that I would change (will change) is the switch as the re is a loud "pop" at switch on. I have been discussing this with someone far more knowledgeable than myself, and adding caps between switch output terminals and voltage rails isn't the answer as this transfers the turn on voltage spike to the woofer from the tweeter; the answer is to change the switch from a single throw double pole to a double throw double pole to reduce the impact to a short "snap" which shouldn't hurt the tweeters. Other than this, it works perfectly. The diagram doesn't show the LED and accompanying resistor as I added that at the build stage.

[StanleyB]

I got quite a few 5532 opamps. So anyone wishing to try out this buffer, or looking to build any other DIY projects with opamps, can have some for free. I just need the postal cost paid. Or you can collect in person.

[sq225917]

http://sjostromaudio.com/pages/index.php/hifi-projects/141-ssb01-sjoestroem-super-buffer-diamond-buffer-free-samples?showall=1&limitstart=

One of the better diamond buffer designs.

[Reffc]

What are the benefits/advantages of that design over the simple 5532 based option Simon? Must admit the only flaw I have found with the one I built was the start up/shut down crack through the speakers (needs a few more components and double pole double throw switch to sort it out).

[DSJR]

Paul, when you can, have a listen to the current spec Albarry preamp. I found the one I last heard eighteen months or so ago just a little too dry and over-disciplined with the matching power amps, but two sources, one here and t'other hifi dave, have told me the latest version is much better and obviously designed from the start to drive very low input-impedance power amps.

I know I said before that we tried the Croft 25RS into the Albarry's and I liked the results very much, but the interconnect was very short and I believe, low capacitance. Others tried this pairing and didn't care for it overmuch. My ancient Croft 4PP preamp also doesn't care for remote power amps, even if their input impedance is around 15K rather than 1.5k, the sound softening and seemingly lacking in 'drive', although into the kind of amps it was designed for (47k upwards), it sounds lovely and very faithful on input-to-output comparisons!

[Reffc]

Paul, when you can, have a listen to the current spec Albarry preamp. I found the one I last heard eighteen months or so ago just a little too dry and over-disciplined with the matching power amps, but two sources, one here and t'other hifi dave, have told me the latest version is much better and obviously designed from the start to drive very low input-impedance power amps.

I know I said before that we tried the Croft 25RS into the Albarry's and I liked the results very much, but the interconnect was very short and I believe, low capacitance. Others tried this pairing and didn't care for it overmuch. My ancient Croft 4PP preamp also doesn't care for remote power amps, even if their input impedance is around 15K rather than 1.5k, the sound softening and seemingly lacking in 'drive', although into the kind of amps it was designed for (47k upwards), it sounds lovely and very faithful on input-to-output comparisons!

Problem is David with any valve pre, the stated output impedance value cannot be reliably used as a guide (ie taking the usual 10x figure for target minimum input impedance for the power amps) because most valve preamps are capacitor coupled, and output impedance varies with frequency. Stated figures are usually for 1KHz but by 10KHz, output impedance could well be 1KOhm or more, hence they really need to be matched with high input impedance power amps. the 7 and 7R are I believe 470kOhms so they present an easy load for their (matching) preamps. I'd say that the Croft really needs a minimum of around 20 kOhms as a load if ICs etc are not to have a noticeable effect at a metre or two.

I don't know about when the latest Albarry preamp was designed. I suspect, like the M608s, the "latest" spec means the AP11 (purpose designed to match the new M608 and M1108 power amps) with onboard phono stage which has been going a few years now. As chance would have it, I've just bought a used one which is about 2 years old so guess that this is the later spec. It arrives tomorrow so I'll do a direct comparison with the Croft and report back.

[struth]

I have a Yaqin valve jobbie; i quite like it but have to say it is the only one i have tried and now use it between cd and ss amp just for the valve effect....my questin is are these valve jobs crap or just different??

[Reffc]

I have a Yaqin valve jobbie; i quite like it but have to say it is the only one i have tried and now use it between cd and ss amp just for the valve effect....my questin is are these valve jobs crap or just different??

If it's the Yaqin valve buffer, I think it uses a basic cathode follower circuit. Benefits sonically are entirely dependant upon which bits of kit it's used between as like all other items of kit, it's part of an overall electrical circuit. The main issue with many valve buffers is that when placed in between SS components with the aim of getting that "valve sound" they also introduce noise meaning a lowering of S/N ratio = lowering fidelity. What ultimately matters though is whether you prefer the sound with it in the system or not, and not what anyone else thinks. If you like what it does, great. If it's subtle at best, one has to ask whether its worth having in the system as it's unusual to have impedance mis-matches between source and preamp unless the preamp uses a particularly low input impedance (low value pot for example).

[struth]

yep, I like listening to 60s/70s stuff through it....a lot of that stuff sounds too sterile on cd I feel....I have the dig out going to a dac as well so i get the best of both worlds though.

[Reffc]

Well, I have the Ap11 pre-amp here and I have to say that its less "lean" sounding than the Croft, probably as it's a lower 110 Ohm output impedance. You were right David, in that the upgrades to the preamp were carried out last year, so this one pre-dates the updates. I have to say though that I don't find it lean. Sound is fuller and more neutral (even warm side of neutral?) than with the Croft in place. Superb S/N (really low noise floor). Very impressed with the match.

[Arkless Electronics]

If powering this of batteries especially I would add a capacitor of at least 100uF to ground from the + and - power supply rails ;)

[Reffc]

If powering this of batteries especially I would add a capacitor of at least 100uF to ground from the + and - power supply rails ;)

Hi Jez

that was done but simply transferred the HF crack to the woofer with a longer duration "thump"! Bit alarming really. Double throw switch with caps across rails and a few other tweaks seems sensible...Irony is that its out of the system as the Albarry AP11 is now in so no need of the buffer now...still, an interesting exercise!

[Arkless Electronics]

Hi Jez

that was done but simply transferred the HF crack to the woofer with a longer duration "thump"! Bit alarming really. Double throw switch with caps across rails and a few other tweaks seems sensible...Irony is that its out of the system as the Albarry AP11 is now in so no need of the buffer now...still, an interesting exercise!

Yer switch the power amp on last! The caps will vastly reduce the PSU impedance at higher frequencies where the op amps PSRR is running out and help compensate for the rising internal resistance of the batteries as they run down....

[Reffc]

Power amp tends to stay on most of the time...point being its not a good solution if I have to keep switching things off and on...(I'd rather design out the problem than treat the symptom) but yes, the simple answer is to ensure that the power amp is switched off first. I have a couple of 100uF caps in my parts bin somewhere I'm sure but not needing the buffer now means I'll probably not get round to fitting them.

[Barry]

Power amp tends to stay on most of the time...point being its not a good solution if I have to keep switching things off and on...(I'd rather design out the problem than treat the symptom) but yes, the simple answer is to ensure that the power amp is switched off first. I have a couple of 100uF caps in my parts bin somewhere I'm sure but not needing the buffer now means I'll probably not get round to fitting them.

You are a brave man, either that or you can afford to be profligate with electricity! I keep my preamps and ESLs powered up all the time, but the power amps (which are 25W Class A designs) consume 800 W, so they are only switched on when needed. Everything, apart from the ESLs, is switched off when I'm out of the country.

[Reffc]

Mine are Class A/B Barry and taking into consideration power factor when not in use, they consume 8W on standby...about the same as a low energy light bulb.