As I mentioned this is one of those that might be different at the detail level in parasitic capacitance due to layout, orientation, the DC specs will be the same.
Had to go to the PO to pick up Robert's pre-amp this morning. Plugged it in and let it warm up for a few hours, listened to side A of Vangelis' "Earth" Vertigo BT-8019 Japanese pressing, "Love Over Gold" side 1, and an EMI Berglund version of Sibelius' "En Saga", everything very nice (VERY nice) I have no initial problems whatsoever. The question will become does this approach have a distinct "voice" as compared to digital RIAA or any other approach or are they all when done carefully nearly indistinguishable.
EDIT - Not too bad on Test Dept either hmmm.... 🙂
EDIT - Not too bad on Test Dept either hmmm.... 🙂
Last edited:
Demain, you might have gotten away with it, but Ed Oxner told me that there was a small asymmetry and at RF frequencies it might be noticeable. I think that he was referring to the NC geometry that the 2N5564 was made from at Siliconix. I should have referred to the Siliconix J113. Of course it is apparently only parasitic capacitance differences between source and gate. There is confusion with the 2N5564 which National uses process 96, not 2 X process 51 like Siliconix did.
Last edited:
Do you have a static THD number at 1K? I haven't measured it in a long time. Lots of incremental changes since the published THD.
BTW Break-in is about 20 hours.
Last edited:
I'm already building my you-know-what off. Herb in Stereophile mentioned it in a recent TT review. I can hardly keep up. Not much time to even post here on DIYAudio 😱
I need to setup ARTA and do some measurements. There is nothing in the signal path but a JC-2 clone (I honestly don't think it uses any fake parts) directly driving 300 Ohm Sennheiser's. Also no facility to do anything but sighted listening. FWIW I have a hard time convincing myself of a consistent difference compared to a standard 5534 feedback RIAA, in this case powered by a +-15V lab brick, which is usual for me.
Really, big TO outputs and 12 O'clock is more than loud enough. I don't listen to anything at loud audiophile levels.
I started with 50ma output Iq and T0-5 transistors with heatsinks for 600 ohms, for the GD, but the Levinson JC-2 preamp used T0-92 transistors, so I was wondering. For the record, the design can do it, but it has to stay Class A, and loading reduces open loop gain proportionally, because the circuit is not an OP-AMP, but a transconductance Amp.
I noticed that, this is the line stage only and TO devices are used in the output maybe they wanted to make it drive 600 Ohms no problem that would mean the 300 Ohm headphones at lowish level should be OK. It does sound fine to my old ears, I will investigate maybe I made a bad assumption.
So I guess none of the nerds here want an electronic beam furnace with 25kv supply. It is BYOO, bring your own oil, for the transformers.
Good, the source and drain are the diagonal rows of dots and the channels are the tiny streets between them (100's effectively in parallel.). The gate is the backside/paddle. Every channel that does not have a dot on both sides could have a VERY small difference in parasitics and there are fractional pF's difference in how the S and D leads are bonded/dressed. You will notice each piece of channel with a dot on each side is a text book symmetrical FET structure. The gold square in the lower left is around 100 microns across give or take a little.
Last edited:
The same idea was used in high speed MOS digital circuits, to minimize polysilicon gate resistance and also to reduce junction perimeter capacitance. Here it is in an ISSCC 1981 paper on a fast memory chip. Figure below.
_
_
Attachments
Reminds me of my time with power mosfets! Always driving gate resistance down and gate capacitance. These were for power switching applications - so a different set of requirements for linear devices and of course JFET's
There are usually fifty or a hundred hand-crafted, lovingly polished transistor devices on a high speed "digital" chip. The highest GBW voltage comparators (which digital guys call "sense amps") have their diff pairs and PFB-coupled load devices built with cross-quad partitioning, dummy device surrounds to cancel proximity effects, even numbers of gate fingers, large area/perimeter ratio, maximum contact density, numerous well taps, and copious excess metallization. Similar amounts of lavish device optimization are applied in the clock trees, serial I/O, and reference generators.
Basket weave layouts (for lateral FETs on ICs) came and went, when people started doing 2D field modeling, and discovered that basket weave has higher source resistance per unit width, than plain old 2N-finger "trombone" layout with machine gun contacts. Then weaves returned when salicided source/drains and vias-stacked-over-contacts arrived. Then they disappeared again when metal-1 resistance rose. And now that finFET layout rules prevent all but the most rectilinear layouts, I think they might be gone for good.
Basket weave layouts (for lateral FETs on ICs) came and went, when people started doing 2D field modeling, and discovered that basket weave has higher source resistance per unit width, than plain old 2N-finger "trombone" layout with machine gun contacts. Then weaves returned when salicided source/drains and vias-stacked-over-contacts arrived. Then they disappeared again when metal-1 resistance rose. And now that finFET layout rules prevent all but the most rectilinear layouts, I think they might be gone for good.
Last edited:
I went to the AA meeting of linear designers on Sunday, when everybody else was watching the big game. Most of you, who sometimes attend, were not there, but I had a very good time talking in depth with Kieth Johnson, who I consider a true competitor, and colleague; much like Charles Hansen. There were others as well, including Kirkwood Rough, who has contributed here, and who has a new power amp design that is an 'engineer's delight' being almost impossible to understand well without extra thought. Also, my old friend Ron Quan was there, the guy who still measures NEW distortions that others tend to ignore. We had a good time, with the big game up on big screens everywhere, which most of us appeared to ignore for the most part. I am glad that I went.
- Status
- Not open for further replies.