[ppl]

You could try the 2SK389 as i discribed above using a battery and see what kind of current you get. The Isolation resistor showen in my attached schematic will isolate that kind of Capacitence and Cascoding will also reduce it alot futher but it is always best to start with a low capacitence device to begin with. As far as the circuit you posted in the gif attachment see my post on walt jungs Multiloop I think it will be alot better as the link will inlighten as to why.

 

[transducer]

Thanks ppl. I looked at Walt's article, actually saw it years ago, but it is well worth revisiting. I will see about implementing his suggestions into the gif schematic that I posted. Really want to try the THS6012, it's spec's are amazing. Will use the OPA627BP biased into Class A. Might also try a trio of Buf634's.

Cheers,
transducer

 

[Possum]

How do you troubleshoot something like this? I ended up using MPF 102's with a 61.9 Ohm resistor between the outputs of an AD823 and the -9V rail. I get sound, but it's very distorted and fuzzy. I've double-checked the MPF 102 datasheets to make sure I used the correct drain, source, and gate pins.

 

[aos]

Measure the current it's drawing. That's the voltage accross your 61.9 Ohm resistor divided by 61.9. It should be a couple of mA, something like 2 or 3.

 

[Possum]

Ah, thanks, it's working now. Current draw is ~3.5 mA. Problem was caused by some parts of the transistors lightly touching parts of my circuitboard. I had to put the transistors and resistor setup on the underside of my PCB since the layout was planned without these extra parts in mind, leaving no room on top.

 

[aos]

So, did you notice improvement in sound?

 

[Possum]

Err, I guess that's what separates the "audiophiles" from the circuit makers/tweakers.

I honestly wish I could A/B the sound, but as long as the theory behind the aspects of a circuit or a tweak sound reasonable, then I'm satisfied if whatever I hear coming out is enjoyable. I guess I actually would A/B it if I wasn't in the middle of cramming for an important exam I have tomorrow. I took a break from studying because I just had to try this out (my Digikey parts came 1 day early - tracking number was expecting delivery tomorrow).

 

[aos]

2N5484 gives only about 1.7mA. I'd prefer about double of that. MPF102 looks like a more fitting choice here.

 

[ppl]

aos: 1.7 Ma. is about right for an opamp driving a high impedance load like a Buffer. I have 1.2Ma using a selected LS-840 with 0.75 Ma. Typical of the LS-840 in cascode. Years ago when i first investigated Biasing Opamps into Class A I found most of The App Notes where this was used (See Data sheets for LT-1115,EL-2008, EL-2009 For starters) were Biasing at 1/10th the Nominal rated Outtput Curent of the Device. After lots of Phone conversations with app Eng's. from several Semiconductor Manufactures Thay all seemed to sugest that this 1/10 was the Max Current and would not Recomend Going Higher. My thinking on using less is the fact that if the Opamp is driving a load of several Meg's as is the case with an Opamp Buffer Combo. The Opamp is not going to see anywhere close to a Ma. of output Current. So why Over-Bias it.

My AB comparrisons on this were easy for me to do because i put the resistor that goses from the Current source to the Opamp in Those machiened IC socket Pins so as i can just pull out the resistor to be able to to see what it is like without it. and i can do this while the Circuit is still Powered. Although Doing AB testing was not the reason for wanting to disable the Current source it did help alot in AB evaluations. The reason i wanted to disable the Current source is to be able to use the AD-744 in it's best sounding mode when Driving a Buffer and that is to use Pin#5 rather than Pin#6 as the Opamps Output. This bypasses the AD-744's internal Output stage and allows one to Connect the VAS stage of the Opamp Direct to the Buffers input and thus avoiding the Added Distortions and time delays any output stage adds. The AD-744 operated in this way dose not require a Current source to Operate into class A since the VAS stage of the Opamp is allready in Class A. The sound Change from operating the 744 in this mode vs using the Opamps internal output stage Biased with a current source or not is quite Impressive. I now never use the AD-744 in a conventional way when it's driving a Buffer. WARNING do not do this without a buffer between the AD-744 and the load since only about 2 ma of current is available and you want to Keep capacitence low so as not to lose Bandwidth. Remember Pin 5 is at a High Impedance Node. Using the AD-744 in this way and driving a buffer results in oune of the Best overall warm and musical Configuations I have Heared. While other opamps may provide more detail or added Brightness. Maby more Bass Slam with some, I think that for Overall Musicality the AD-744 Driving a buffer operated with Pin#5 as the Output of the AD-744 is a hard one to beat. and it seems that no matter what Opamp i try that after awhile i always go back to Silky Smooth Euphoric sound of the AD-744 using pin#5 as stated above. Opamps like the AD-825 and others provide better Detail and slightly better sound stage but also force you to Hear the Bad Recording Job done on most recordings and this is starting to Irritate me. I may make this into another thred as it makes for an easy search in the future.

 

[Tomo]

Hello,

I am raising this discussion back into our focus.

TOMO

 

[Joobu]

Finally got around to reading some more app notes. The Linear LT1575 is an ultrafast low dropout regulator. Designed to drive microprocessors with large transient currents on the order of nano seconds.

Basically it's a linear regulator with a huge amount of bandwidth. Due to this, designs can do away with the huge bulk decoupling usually employed. Of course you still need some local ceramic decoupling. Nonetheless, definitely seems like a better approach than putting down hundreds of thousand of farads of capacitance

 

[ppl]

Joobu; LTC has been saying the same thing for years starting with the old LT-1085 & 1086. However look at any Mother-Board next time your at a PC shop and you will see Big Electro's Bypassing whatever Regulator is used. Years ago I used to think as you did Thanks to mark brandsfield of the Then MSB technology but after lots of listening over the Years i have Found Big caps Sound Better will smooth any Glitch from the reg's. All Regs regardless of Specs have Ringing on fast High Current pulse. Granted the magnatude is small but the Big cap reducesit Even futher. Sound Quality is also improved. This is only My View and other views with equal technical Support are available from other sources and some on these fourms. So Ya all save your Flames

 

[Joobu]

You'll always have bulk decoupling. However, I wonder how much of that bulk decoupling is actually powering the processor. Power circuitry for processors is an involved design. Step-down switchers (I'm pretty sure that's what all those toroids are for) and of course linear regulators.
I'm never tried the regulator so I don't know how valid its merits are but it's definitely an interesting implementation.

 

[ppl]

Joobu: Power supply design for High Current High data rate Digital Circuits are in a different class than that of Low Noise Analog Circuits. In Analog design switchers are often not Recomended, However thay still are used in some situations. aos and tomo might be the ones to eloberate upon Fast regulators.

 

[aos]

I've read tons of datasheets for regulators, but I looked only at linear regulators. Some very fast linear regulators will regulate well in the MHz region and are usable for digital circuits. The ones I use are meant to be used for Bluetooth and handheld devices. The voltages there are usually so low that I guess they don't need switching to up the voltage.

I'd love to see a nice article where someone professional, with profi equipment naturally, has done comparison of vanilla LM317/337/7815 etc. with these new regulators. But I would encourage people to look beyond those old 3-pin regulators.