(Schematic)

What will not be changing:

- Serpac H-65 compatible board size
- ALPS RK27 and RK097/8 support
- Thru-hole where possible
- Jung multiloop topology, with class A biased op-amp

What has changed:

- 4-layer, with near-complete ground plane! (Bottom layer, blue above, does have some traces running through it in the few cases where running them in the top layer isn't possible work.)

- Thicker traces in several places

- Switched from BUF634 to LMH6321. The availability question is forcing the issue, but these chips sound a lot better, even when you compare to BUF634s biased to draw the same current. There are downsides: requires SMT, can't get down to lower supply currents for long-running amps, may be tying ourselves into another chip that will go away.

- Main rail cap bank goes from rail to rail only, like on PPA. This puts more burden on the ground channel to maintain low distortion. On the plus side, it allows the ground channel more freedom to do this, without getting swamped. It also increases the effective size of the rail cap bank, even though I had to drop 1 of them to achieve other goals.

- Bypass caps at op-amp pins, with choice of either PCM-5 thru-hole or SMT 1206 style packages.

- More mounting holes. (Unfortunate consequence: the outer holes aren't 1.4" apart any more. To mount a crossfeed board, you have to use holes 1 and 3, or 2 and 4.)

- Front shoulders moved back, inner corners rounded, notch added in rear edge, and main cap bank moved in slightly to make board fit better in Serpac H-65 enclosure

- Output pads are no longer clustered. My philosophy of putting the pad nearest the last place in the circuit where the signal naturally exists was being violated in PIMETA v1.

- Several other pad sets are clustered better now, where it doesn't hurt. Allowing room for Molex KK 0.1" pitch connectors in these places.

- Both ALPS pots are centered now, not just RK097

- Dropped Panasonic EVJ-C footprint. I have seen hardly any of these being used since I started carrying the RK097s. If I could have kept it without difficulties in routing, I might have, but since it costs, it goes.

- PG pad placement more useful for ALPS RK27.

- Bigger C1s allowed. (Still jumpered by default.)

- Only one set of input and switch pads. (PIMETA v1.1 used separate ones for RK097 pot case, a big source of confusion.)

What's still up in the air:

- Keep the 4-layer design? This will double the board cost.

- Dropped R8. (Output resistor inside feedback loop.) I can't remember the last time I saw this used at all, or recommended adding it. Unless someone comes up with reasons it's necessary, backed by actual experience with PIMETA v1, it stays out. Note that it's still possible to add an output resistor outside the loop, just by putting one between each O pad and the output jack.

- C4s are no longer on the center line. I can't seem to work out a layout that allows this, without having worse consequences than mere esthetics.

- The current layer stack-up doesn't entirely make me happy. On the plus side, it keeps all traces on the outer layers so you can follow then without looking at CAD drawings. On the downside, it doesn't put ground between the power planes as makes intuitive sense, and I'd prefer V- to be on top to allow better heat sinking for the buffers. That suggests a stack-up of V-, GND, signal, V+, but then you're blind to the traces. Not sure which option makes more sense here...

- Bring back low-voltage LED cut-off? I haven't seen many people use it, so I dropped it. Conflicts with one of the new CCS options, too.

- Class A biasing CCS type. The current one has a number of problems:

-- requires buying lots of JFETs to get a particular current level, and each one has to be hand-tested

-- consequently, almost no one does this, requiring me to do the matching, which for efficiency reasons still has a lot of drawbacks

-- once you select the current, changing it requires going clear back through the process, and replacing the current CCS

-- some say it doesn't perform as well as other possibilities

So, I'm looking at other circuits:

LTspice says they have this dropout behavior:

(LTspice IV file)

Given how soggy the bipolar ones are compared to the JFET ones, you might be wondering why I'm even talking about this option, much less have bothered to do the layout for one of them. They have several merits:

- The first one uses the fact that you almost certainly want a power LED, using it as a voltage reference to set the bias point for the main CCS. That current is set by a single resistor, and then gets mirrored twice to tug on the left and right channels. This means the two channels' bias points are always matched, and can be changed with just one resistor, currently a trim pot with series resistance to set a maximum bias of around 10 mA. The main downside of this option is that there's a waste path through Q2, equal to the op-amp bias current: you multiply the bias current by 3, not 2. This is still better than a naive implementation of this CCS, which would have an LED per channel, and two waste paths, plus the need to bias both channels separately.

- The second one trades performance for less waste current by taking advantage of the fact that we typically bias op-amps with around the same amount of current as we drive LEDs. By using diodes instead of the power LED to bias the main CCS and moving the LED into that Q2 waste path, we make the Q2 path do real work, while allowing less current use in the main CCS biasing path. This is likely to outperform the PIMETA v1.1 current draw when comparing the random JFETs you can get from my shop now, as those tend toward higher currents than you might choose if you had the power. We also get constant LED brightness with batteries as a nice side effect. The main downside is that the LED Vf increases the dropout voltage. Note that in the graph above, the diode-biased CCS cuts off just a little above 3.6 V, that being the Vf of the LED I'm using in the simulation. This is a minor matter, though, given that the amp as a whole isn't likely to be able to get down to this level.

Comments, questions, complaints?

---

[Quoting tangent]

- Bypass caps at op-amp pins, with choice of either PCM-5 thru-hole or SMT 1206 style packages.

---

My limited opamp experience says that this makes a big deal. Nice to see them added.

Another "up in the air" matter I forgot to mention: the LMH6321 isn't current-limited by default. You can add a resistor to do this, which would protect against people pulling their headphone plugs out while music is playing. We haven't seen a lot of trouble with this on the PPA, which uses buffers with no output current limiting, but then, I imagine almost all PPAs use 1/4" output jacks, where this is harder to do. A lot of PIMETAs will use 1/8" jacks. Should I add resistors for this? It might be tricky to get them in there, but the pain might be worth it.

---

[Quoting tangent]

- Switched from BUF634 to LMH6321. The availability question is forcing the issue, but these chips sound a lot better, even when you compare to BUF634s biased to draw the same current. There are downsides: requires SMT

---

I am still mostly a beginner in DIY -- may I just say that I support the use of SMT here. This is the perfect place to introduce and encourage SMT soldering, but for more than just SMT's sake.

---

[Quoting tangent]

Should I add resistors for this? It might be tricky to get them in there, but the pain might be worth it.

---

My sense would be that with spots for them, people can jumper. Without spots, then they have no option which could lead to issues down the road. If you don't put them in, then you might want to put the buffers on adapters so that they are easily replaceable.

Didn't someone else (Linear, National?) start producing buf634's with a different name? If so, maybe DIP-8 spots for the buffers, and custom browndog-like adapter boards for using the different buffers. This also makes it easier for someone to drop in a discrete buffer, which is bound to happen.

This may well be flawed models. Many BJT models are pessimistic about Early Effect. Many JFET models do not give honest conductances. It may be more like one order of magnitude, not two orders as your models are saying. (And some JFETs have very much lower plate resistance than these models give.)

And the AC impedance may be quite different from DC readings. And real hard to measure meaningfully.

Your "worst" is 18K DC and 180K AC. Relative to the ~~28R impedance of an AB BJT output stage, this is "very high", I would not sneer on this basis alone.

Worst dropout is 4V; as you say, I bet few 9V-18V systems will still be happy as V1 falls to 4V.

The Diode-bias has two points: the base-bias current is quite small, starving the 1N914s; and you tucked an LED into the collector which gives large dropout. It would "look better" with R7=3.3K and the LED taken away. (Note that you could reduce 3.3K to say 2.7K and put the LED in this loop.)

Emitter resistors under the 2N3904s might bump the plate resistance; also unless you are getting very well-matched '3904s then another 20mV-100mV of dead resistance might improve mirror consistency.

> a naive implementation of this CCS, which would have an LED per channel, and two waste paths

Use Red (1.6V) LED. Nail to negative rail, pull-up with say 3K or 10K. Two 2N3904 with ~~1K emitter resistors will pull-down 1mA each channel. Pull-down won't happen when node is within ~~1V of rail; but the main thing is to clean the small-signal 9V up from rail, not to stay perfect to within a half-dB of clipping.

Personally, I'd throw a dart.... but that would let the magic juice out of my PC's LCD display....

---

[Quoting cobaltmute]

What about bypass on the buffers?

---

If there weren't a real estate cost, the decision might have gone differently. As it is, I'm talking myself into using what little real estate there is for CL resistors.

---

[Quoting dsavitsk]

My sense would be that with spots for them, people can jumper.

---

---

[Quote]

If you don't put them in, then you might want to put the buffers on adapters so that they are easily replaceable.

---

As for replacement, it's actually not that difficult to remove an SO-8 chip from a PCB. You don't need anything but a soldering iron to do it. I show how in Tangent Tutorial 4.

---

[Quote]

Didn't someone else (Linear, National?) start producing buf634's with a different name?

---

---

[Quote]

What about using your Pint setup for the rail splitter?

---

When you're not so forced, you might choose not to pay the costs:

- Another op-amp channel is generally more expensive than a TLE2426

- The extra op-amp draws a lot more current than a TLE2426

- The two external resistors you have to add take more board space than a TO-92, unless you tombstone them

---

[Quote]

thus doing away with the TLE once and for all.

---

---

[Quoting PRR]

This may well be flawed models.

---

---

[Quote]

And the AC impedance may be quite different from DC readings. And real hard to measure meaningfully.

---

I'm pretty sure SPICE can be made to do the right thing -- 18 VDC + 1 VAC ripple -- but I gave up on it. I attached the circuit file above, so if someone gives it a go and figures it out, please post it.

---

[Quote]

The Diode-bias has two points: the base-bias current is quite small, starving the 1N914s

---

That brings me to option 3, which I just thought of today: put pads on the board to allow an LED in either position. You then have several options in populating the board. You can configure it like either bipolar CCS above, by suitable use of jumpers and strings of diodes in a loop in the LED pads. Or, you can use both, giving more interior illumination for a clear enclosure, or just to get better performance.

---

[Quote]

Emitter resistors under the 2N3904s might bump the plate resistance; also unless you are getting very well-matched '3904s then another 20mV-100mV of dead resistance might improve mirror consistency.

---

---

[Quoting tangent]

Aside from availability difficulties, is there anything you see wrong with the TLE2426, as compared to an op-amp and a pair of resistors?

---

They seem like a common point of failure. I burned one up in an otherwise happy Pimeta a long time ago, and I've seen them fail in other places, too. I think that this is not that uncommon, though I could be wrong.

I do not know. Long ago, when I cared, it was not uncommon for "identical" BJT parts to be 20mV off at the same current, implying 2:1 off in current at the same Vbe. But I think wafer-fab has improved a ton, and on-tape parts will generally be same-wafer and probably adjacent dice, therefore very much the same. Others report +/-2mV deviation on BJTs from a reel.

I put some coal under my SPICE to see what it says. Yes, it says a simple JFET is much better than a couple BJTs. The problem is variability; general-purpose JFETs may be 2mA or 8mA Idss, and similar range of Vp, meaning 3:1 or 4:1 range of actual current. Which maybe won't bother the op-amp, but is sure to upset some fussy builders.

There are for-purpose Idss-selected JFETs. DigiKey can't find any, Newark does not stock, Mouser does. But Mouser's in-stock parts start at $1.02, and the good-range parts are $1.66 for +/-20% selection. Not a budget-breaker but not trivial. But this is a one-part no-fiddle solution to off-biasing the opamp chip, so may be "cheaper" in some way than a collection of $0.29 parts and PCB space or trims.

Part Number Min _ Nom _ Max _Zd
SST504 ___ 0.600 0.75 0.900 0.5Meg
SST505 ___ 0.800 1.00 1.200 0.4
SST506 ___ 1.120 1.40 1.680 0.3
http://www.visha...70711/70711.pdf

FWIW: the "Typical" dynamic impedance (not shown) is mighty near what SPICE gives for a JFET plus a 2K resistor to set current. The Min (shown) is far higher than we need here.

+/-10% parts are $34!!! At this point you can buy a $13 DVM and a box of resistors to trim with.

I understood that. I too resent the "waste". But I'm not coming up with any way to use that waste, not which _I_ like. I have no real objection to any of your plans, and no real preference. I sense you don't have a favorite either. Either we both dumb, or there isn't a "great!" solution.

What we "obviously" want is a way to get ~~2V at set current from 18V-10V with "no" waste. A rectangle wave with 10% duty cycle can be very low voltage waste. Add a coil and adjust frequency to get low current waste. Same as a Switching Voltage-Regulator, only different. And awful small for the ready-made switcher chips. And major designer-distraction, when the focus should be on the audio.

If you would allow the pull-down CSSes to drop-out on huge negative swings: LED up from Neg rail. Two CSS JFETs from opamps to LED. 1mA pull-down per channel gives 2mA LED current. Assuming 2V LED and 1V FET and 9V rail, pull-down works to 6V peak, "fails" for the largest peaks, but we surely clip by 8V or 7V. If a little more signal gives total loss of accuracy, I'm not inclined to care if the pull-down "only" works to 6V, and after that there may be an A-B crossover. In real speech/music this should be a very rare event, unless you have pushed into frequent clipping then so what?

In extreme clipping, the LED could dim; so what?

I'm quite sure battery life will be better with LARGE caps, either in/out or if you must pretend "capacitorless", then swamping the rail-splitter. With the pure active rail-splitter, most actual signal power is dissipated in the splitter. With ample caps, this power is re-cycled to the next half-wave.

I preached that before and got no takers. The experiment seems too simple to avoid. Devise a consistent source, such as a CD on Repeat. Buy a 10-pack of 9V batts and select four of equal voltage. Load an amp, play until batteries sag to say 8.0V. Note time. Then build-out the rail-splitter with 100R, add two 1,000uFd rail-caps, change batteries, repeat the run. I predict the splitter will run cooler and the batts will last longer.

Yes, 2x1,000uFd is "big". Less will work. And if indeed the batts last longer, is that worth something? Maybe a smaller battery, or not carrying spares?

Bass crosstalk will rise. There isn't a great answer for that. Except wiring single-ended with output coupling caps. And that violates much dogma. Or 4-wire headphones, which is not generally done.

I realize you are doing an incremental improvement, not a radical re-think. I'm just jabbering with post-Flu false energy.

Tangent: when my PPAv2 was working I used an 1/8 inch jack and never had any trouble when unplugging my SE530's from it.

[Edited by fsma on 02-10-2009 at 02:01 AM.]

Could I use this opportunity to ask a slightly irrelevant question: Does the LMH6321 have a DC offset on its output? Its datasheet is unclear.

Not an issue with this circuitm but I'm thinking of using this in an open-loop application (where the BUF634 is not cool due to its very large DC offset).

Sorry for the threadjack. You may now return to your regularly scheduled redesign.

---

[Quoting PRR]

I'm quite sure battery life will be better with LARGE caps, either in/out or if you must pretend "capacitorless", then swamping the rail-splitter. With the pure active rail-splitter, most actual signal power is dissipated in the splitter. With ample caps, this power is re-cycled to the next half-wave.

---

---

[Quote]

I'm just jabbering with post-Flu false energy.

---

---

[Quoting dougigs]

Does the LMH6321 have a DC offset on its output?

---

---

[Quote]

Its datasheet is unclear.

---

About the TLE - it became unobtainium for a few months ... not nearly as bad as the BUF, but it could happen again.

I think SMD is great as an option, but the PIMETA has always been a 1st choice after the CMoy, IMHO. Having some parts requiring SMD might raise the difficulty level beyond post-CMoy capability. If that's what it takes to solve the buffer availability, though, then maybe it's justified.

P.S. Wasn't someone using the AD744 as a buffer with the compensation pin lifted - or something like that?

[Edited by tomb on 02-10-2009 at 03:36 PM.]

1) offboard molex: good I like that I can ignore the pads for the pot and short each wiper to the non-ground end and then use ONLY the molex to feed the circuit's true input. for people like me who use the pimeta board as a small footprint preamp module, I don't like having to tap into the pot pads for inputs; a 3 pin molex is nice and clean.

2) I'd really like 3 ADJACENT pads for output, too. I realize you don't want to do the routing (or can't) and you like having the pads near the output points, as close as possible to the circuit points, but if you provide 3 'landing pads' (just isolated pads in molex .1 centers, somewhere) then I can at least mount a molex 3block there and tap into the landing pads with greenwires and get at least a physical block to connect to.

3) bass boost; I like having it - it adds a nice bit of free warmth why not have some pads that let the user do at least a 2 level or 3 level switched boost? I like the idea of tiny relays doing it (like my daughtercard hack) and it sure would be nice (for non-portable builds) to be able to relay switch a few levels of bass boost.

4) crossfeed; is there no way to fit crossfeed on the board? would going SMD help? maybe a small optional area on the board for an smd-only crossfeed section?

5) dc-dc conversion; having to run on 9v or higher and also not deriving a minus is sort of a down side. I realize that charge pumps and such have a bad rep in the high end circles but being able to run on, say, 2 AA's would be a HUGE win.

6) turn-on delay; any way to add SOMETHING, even if not a full proper solution, to stop turn on/off thuds? solid state mux chip for analog switch? I really hate turn on/off thuds. fixing that would bring this to a whole new level (I already feel the audio is 'fine enough' and its mostly the muting that needs help in amps and dacs, imho!)

7) 4layer board; this adds to cost and makes hacking the board harder. also makes home etching impossible. I'm not sure its worth it to go 4layer.. in a related topic, please considering making the voids around the traces (the ground plane 'surround') having wider gaps than what is shown. it makes visual inspection easier if the gaps are not so tiny, it also makes bridges less likely and also is a bit more forgiving to manufacture! an extra mm isn't going to kill performance and for diy builders, it would be welcome to have a bit more 'breathing room' around traces. we are not pro builders and so this should be taken into account

8 ) any benefit to having a TLE per chip instead of per board?

9) any benefit (or, can the space be justified) by going to single amp chips instead of dual-opamp chips? that also makes troubleshooting easier as, on the ppa, I can swap the opag around with opal and olar and I can easily find the bad part that way. if my stereo chip is bad, I can't swap it and find out, using this same circuit. minor detail but it does help in troubleshooting and sourcing parts.

that's about all I can think of - for now

[Edited by linux-works on 02-10-2009 at 03:59 PM.]

its a big departure from the analog pot currently in place in the v1, but why not consider this for the v2?

I played with an 1802 and kind of liked it - I just never fully finished my version.

when listening to headphones, I often find the 'nudge' model of volume control to be better than the pot. I might need a bit more or less volume and so a 'single click' of a vol control is often all I need. and its nice to know that this keeps the channels perfectly in balance and also 'tests well' as they say, in many aspects

by removing the pot, you also save space and some height, if that matters.

oh, and in terms of height, why not allow for 'option modules' to be daughter cards (of a sort) and have a well defined socket interface below. maybe if there's a clever way to adapt a crossfeed/bass boost daughtercard so that it fits in the standard box but not on the main pcb, so that builders who don't want those options don't have to 'suffer' due to them

---

[Quoting tomb]

Best wishes on the PIMETA V2! This has always been an outstanding DIY project.

---

---

[Quote]

About the TLE - it became unobtainium for a few months

---

---

[Quote]

Having some parts requiring SMD might raise the difficulty level beyond post-CMoy capability.

---

Mind, I'm only talking about relatively large parts here. SO-8s, 1206 and up 2-lead parts, etc. Maybe even the occasional SOT-23. Truly tiny things...TSSOP and such...should indeed be out of the question for beginner projects.

---

[Quote]

Wasn't someone using the AD744 as a buffer with the compensation pin lifted - or something like that?

---

I've never tried this myself, as I'm a bit of an op-amp snob, and like to select them to season a circuit. There aren't that many options for chips with this feature; none in dual-channel versions, which I need for the PIMETA.

---

[Quoting linux-works]

I like that I can ignore the pads for the pot and short each wiper to the non-ground end and then use ONLY the molex to feed the circuit's true input. for people like me who use the pimeta board as a small footprint preamp module, I don't like having to tap into the pot pads for inputs; a 3 pin molex is nice and clean.

---

---

[Quote]

if you provide 3 'landing pads' (just isolated pads in molex .1 centers, somewhere) then I can at least mount a molex 3block there and tap into the landing pads with greenwires and get at least a physical block to connect to.

---

Speaking of, what if I were to turn the tail edge of the board into a short chunk of patterned perfboard? It could give access to the rail and ground layers, plus have other pads for other things.

---

[Quote]

bass boost

---

---

[Quote]

crossfeed; is there no way to fit crossfeed on the board?

---

---

[Quote]

dc-dc conversion

---

---

[Quote]

turn-on delay; any way to add SOMETHING, even if not a full proper solution, to stop turn on/off thuds?

---

---

[Quote]

4layer board; this adds to cost and makes hacking the board harder.

---

Cost: does making the board $15 change whether you'd buy it, given that it should perform better? It's not like I'm adding layers just because I can. I expect to get a real benefit for it.

Hacking traces: as designed above, this shouldn't be a problem. The signal traces are on the outer layers. It does make it more difficult to separate chips from their power supply layers, but why would you do that?

I guess I can take this as an argument against putting the planes on the outer layers and the signals inside.

---

[Quote]

also makes home etching impossible

---

---

[Quote]

please considering making the voids around the traces (the ground plane 'surround') having wider gaps than what is shown.

---

---

[Quote]

it makes visual inspection easier

---

---

[Quote]

it also makes bridges less likely

---

---

[Quote]

a bit more forgiving to manufacture!

---

---

[Quote]

any benefit to having a TLE per chip instead of per board?

---

---

[Quote]

how about the ds1802 volume control up/down pushbutton chip?

---

---

[Quote]

why not consider this for the v2?

---

---

[Quote]

why not allow for 'option modules' to be daughter cards

---

---

[Quoting vixr]

Man, I swear I learn a ton of stuff every time I read a thread Tangent starts...the brain power you guys give off is humbling.

---

[Edited by tangent on 02-10-2009 at 10:04 PM.]

As for the perf-board block at the back, that is a great idea. If you expose V+, gnd, V-, possibly some switch pads, you allow for a great choice in power supply options (for those that don't use your preferred case).

Another idea is to provide a spot to do a small trickle charger. Say a LM317LZ,the required resistor and the required diodes.