Hi Silviu,
If you prefer balanced, it should work fine.
Some other checkpoints:
For MC input, a low-noise JFET front end is recommended, unless you have a transformer already: see attached serving-suggestion, for the front end MC stage, this is what I use.
Please consider how the volume control will work, if the output of the RIAA stage is balanced.
If you prefer balanced, it should work fine.
Some other checkpoints:
For MC input, a low-noise JFET front end is recommended, unless you have a transformer already: see attached serving-suggestion, for the front end MC stage, this is what I use.
Please consider how the volume control will work, if the output of the RIAA stage is balanced.
Attachments
Hi Rod,
Would this front end be overkill?
hybrid folded cascode Gm amp – lab jc
For volume control I'll probably go with a switched auto-transformer... I have to decide on that also.. my system is in a perpetual change...
Silviu
Would this front end be overkill?
hybrid folded cascode Gm amp – lab jc
For volume control I'll probably go with a switched auto-transformer... I have to decide on that also.. my system is in a perpetual change...
Silviu
You could use JC's circuit - but I can't say I like the 25000µF cap on the source resistor. Beware also that the 2SJ74 has gone out of production since the circuit was posted, and as a p-channel part, it is not easy to substitute.
OTOH, You can parallel multiple N-channel JFETs, in the circuit in post #43, if you have an LOMC.
One important point: bipolar transistors as the cascode element (NPN for series cascode, PNP for Shunt cascode) measure better than series cascode with triode as the upstairs device, and to my ears, sound better too. It's all about gm in this position.
OTOH, You can parallel multiple N-channel JFETs, in the circuit in post #43, if you have an LOMC.
One important point: bipolar transistors as the cascode element (NPN for series cascode, PNP for Shunt cascode) measure better than series cascode with triode as the upstairs device, and to my ears, sound better too. It's all about gm in this position.
The cascode output impedance is very high, similar to that of a pentode because the plate of the lower Triode is shielded from voltage changes. Thus the total output impedance of our first stage is close to 6200Ohm. If you have too much gain, just omit the bypassing capacitor C1. This will increase the impedance of the cascode slightly and your total output impedance will be even closer to the ideal 6200Ohm.
I ran the sim with an AC signal and that BJT 2N5401 in a common base config. The emitter input presents a very low input Z, and the output is high Z. Half of the current sources output goes to the tube, the other to the 2N5401. The voltage on the input tube's plate is held (near) constant, as it is being used as a current amplifier, to feed the BJT emitter.
I don't get the idea,
why use a BJT (IV-converter) for an active element in a tube circuit? Why not use a tube as a voltage amp and buffer the filter with a mu-follower?
I don't get the idea,
why use a BJT (IV-converter) for an active element in a tube circuit? Why not use a tube as a voltage amp and buffer the filter with a mu-follower?
In my limited experience i found triode cascodes for low noise audio circuits very unnatural.Best low noise cascode tubes were made for high FM frequency operation where 50/60 hz fillament noise caused by second(cascode) cathode to fillament noise is less important as you can filter out the whole audio spectrum anytime, but using full tube cascodes in audio is really painfull as you need to set a half anode voltage fillament bias for both tube sections and the input tube need to handle a 100...200 volt FK voltage affecting tube's life expectancy and cathode to fillament leakage current as the usual high GM, low noise triodes can't handle that well unless you're using different tubes with different fillaments which throws you into a space management problem for the fillaments wires which by themselves generate noise especially due to having stereo operation, thus four sets of fillament supply wires and high GM tubes also needs higher fillament currents and their wires are twice the thichness, not to mention 8...16 times more anode currents to be filtered as PSRR is also very poor with cascodes .Then if you're going to use 12ax7 which can handle it easily, you can't use passive LCR loading...Pentodes can't handle low impedance LCR networks well either although usual low GM , low noise pentodes have good enough noise properties hence the second device in a cascode or mu follower is much better to be a silicon transistor.
On the other hand using high gain and impedance cascodes with LCR loading doesn't really make sense as cascodes are much more useful for active riaa circuits.
Using ideal Voltage dependent voltage souces with tube simulations isn't really going to be full of meaning unfortunately... Even with ideal loading tubes are tubes...nothing too ideal about them.
I had my own benefits over cons battles years away and opted for the appropriate circuits even when i went original(not my option anymore...) I started to consider bipolar transistors for any mu-follower or cascode device even when keeping the first device a tube as a transistor active load is much better suited for even lower gm tubes offering perfect conditions for any tube to shine.Almost any other option throws you into over designing your circuits making poor use of resources.It's like i'd still choose to make full tube stereo fm radios in 2021...but overdesigning haunts enyone in the diy world from time to time
My explanations migt be also too late in this subject as i see now, but years ago i wished anyone explaining to me why some ideal circuits aren't suited for any task.
Last edited:
I've tried E810F and 6J52P with "best pentode" circuit, sounds pretty good other than a little weak bass maybe due to the 10K:600 transformer (Electra-print 35ma), also built EC8010 with Gyrator hybrid mu-follower to drive LCR600 directly.
Ideally I want to have first triode stage with enough gain (200x) before the 10K:600, wondering if it's workable to insert as the 10K load for shunt/folded cascode stage.
Ideally I want to have first triode stage with enough gain (200x) before the 10K:600, wondering if it's workable to insert as the 10K load for shunt/folded cascode stage.
It's much better to use a source-follower at the output of the Shunt Cascode.
That way, the load can be almost anything you like, so long as the FET's SOA is respected for the case temperature is must operate at.
Choose an N-ch FET with Crss <12pF at the Vds value you can run it to.
If you're looking to save a stage, better wire the e810f as triode and use a transistor(mos-fet or bipolar diamond buffer) as mu-follower.Lower noise, lower output impedance and very stable operation.You can replace the whole bipolar circuit with the right mos-fet but i had no DN2540 .Both my versions were tried with pasive riaa after the first stage not as in simulations.The bipolar version yielded the best results by far and i could never get better result with any other topology.I briefly tried mosfets but i think i got the bias wrong at the time as it apperaed to have some ultrasonic oscillations and very unnatural sound.Ale Moglia had its own version of gyrator...All in all i think gyrators are better and more stable than cascodes.
https://www.diyaudio.com/forums/analogue-source/255149-unu-pnono-riaa-mm-preamp.html#post5485527
https://www.diyaudio.com/forums/ana...v2-phono-riaa-preamplifier-2.html#post5589877
https://www.diyaudio.com/forums/analogue-source/255149-unu-pnono-riaa-mm-preamp.html#post5485527
https://www.diyaudio.com/forums/ana...v2-phono-riaa-preamplifier-2.html#post5589877
No, it's not, it's actually too low...there's no audio cascode circuit that wouldn't have an additional compensation capacitance with higher value than e810f internal capacitances. E810f is a tube working at 250mHz in pentode mode and mu-follower operation in triode mode liniarises it even more than the pentode operation can do it.In my bipolar -d3a mu-follower i got perfect square wave reproduction up to 25khz and undistorted sine wave up to 480khz...if that's not enough...I think audio world is hostage to too much nonsense for many years.If it would be to use riaa in feedback network i'd understand the need for higher speed operation, but that's completely unnecessary if not toxic with pasive riaa.And i used e810f against d3a finding d3a (lower bandwidth) sounding much better while e810f very unstable in triode mode .These tubes were for VHF amplification man...you need to make them slower for stable audio range!
Last edited:
I thought the Cag was around 3pf, equivalent miller capacitance of 200p.
This is fine for MC/SUT but not friendly for lots of MM.
I put ferrite bead and grid/anode stopper.
Very interesting I saw an internal picture of Allnic H-3000 without any prevention.
This is fine for MC/SUT but not friendly for lots of MM.
I put ferrite bead and grid/anode stopper.
Very interesting I saw an internal picture of Allnic H-3000 without any prevention.
Attachments
Last edited:
e810f is very unnatural to audio range in any triode or pentode wiring but feel free to try it first before any theoretical considerations and believe me, 200pf miller capacitance is not a problem for any good mm cart.Don't judge tubes as you judge transitors cause miller effect is not having the same effect in both .That miller capacitance is in series with a rather high resistor value in tubes including input grid noise resistance of the tube itself . Check luxman cl34 and see what kind of resistor it uses at the input while the whole circuit has global feedback thus rather low perceived miller capacitance while in operation yet the bandwidth is very good.You couldn't use a 3k3 input resistor in bipolar transistors to bend that miller capacitance loading although luxman circuit doesn't really need it to be that high ...Theres' actually one single company that used to load the cartridge with 100ohm series resistor to 100pf and nothing else in its most high end silicon phono preamp and that was actually much more practical than having switched multiple input capacitances as turntables usecd to come with their own wirings having the optimum capacitance by default for their chosen cartridge.In the old times manufacturers didn't offer a high quality turntable for any cartridge you could find on the market, they used to offer the whole package: the best tonearm and wiring for a given cartridge so that it can be attached to any phono preamp having a standard 100pf loading capacitance.Nowadays we can have our preffered setup and addapt the cart and tonearm wring the the phono preamp we already have. I tried e810f , e88cc and d3a even with bootstrapped cascodes as in Tektronix nuvistor based CRT oscilloope and nothing really good came out of it.Mu-followers and gyrators rule over it by far.Give it a try!I spent over a year in 2013-2014 trying all sort of different topologies and allan kimmel mu follower with triode strapped D3A came on top both sonically and measurably.Never tried LCR networks as i couldn't afford ones yet i had the best results in passive configurations like that.If you want the same circuit for both mc and mm carts you should probably consider paralleled e810f with no grid and cathode resistor at all which won't really work for mm operation.