My voltage doesn't sag when i monitored it....it stays at 255v even at full volume.
So my thought is wrong that 5.8v will be the maximum input signal that can be handled by the second stage...maybe for the first stage its applicable...so i have to go back on the 6n6p curve again...
any suggestion how to appropriate the the schematics to work at best? i'm planning to add one more led on 6n6p bias. will it be good?
it is the peak voltage input before the grid becomes positive and draws current, but that is another story...
gain is gm x Rl.......gm depends on plate current, Rl is the parallel combination of rp(tubes plate resistance at specified plate voltage, the impedance of the CCS, and the next stage grid leak resistor) all these taken is always lower than plate resistance....
don't know what you mean by best, but it is best to know the limitations of your circuit so you know what to expect...
gain is gm x Rl.......gm depends on plate current, Rl is the parallel combination of rp(tubes plate resistance at specified plate voltage, the impedance of the CCS, and the next stage grid leak resistor) all these taken is always lower than plate resistance....
don't know what you mean by best, but it is best to know the limitations of your circuit so you know what to expect...
okay sir i will try both options....
BTW, if i will put an option for feedback to lessen the gain....do i need to replaced the 6n1p led for resistor (say 390 ohms) to inject the feedbaCk?
will it be possible to put around 47k fb resistor form opt seconday then to the junction of 6n1p cathode and 390 ohms?
BTW, if i will put an option for feedback to lessen the gain....do i need to replaced the 6n1p led for resistor (say 390 ohms) to inject the feedbaCk?
will it be possible to put around 47k fb resistor form opt seconday then to the junction of 6n1p cathode and 390 ohms?
The question is why the LEDs are flashing. The changes in brightness indicate a variable current through the LEDs, when there should be a constant current (the MOSFET load should draw approximately zero current). The voltage across the LEDs looks the same because the impedance is only probably about 30 ohms and a current differential of 5 ma would only produce about 150mv of change and most meters aren't fast enough to see this as music plays.
There's only one place where the missing current can be going - through the grid. The first stage is probably overdriving the grid causing current to flow. The gain of the 6N1P is too high (it was originally my first choice too). I would use a voltage divider on the output of the 6N1P to reduce the signal out. A voltage divider on the 6N6P would shift the DC bias, which is pretty good where it is. Otherwise, the design looks pretty good.
There's only one place where the missing current can be going - through the grid. The first stage is probably overdriving the grid causing current to flow. The gain of the 6N1P is too high (it was originally my first choice too). I would use a voltage divider on the output of the 6N1P to reduce the signal out. A voltage divider on the 6N6P would shift the DC bias, which is pretty good where it is. Otherwise, the design looks pretty good.
this is exactly in my mind....so i'm going to put a voltage divider after the 6N1P to see..what will be the effect...and also replacing the 6n6p CCS source resistor with trimpot so that i can adjust the 1/2 B+...
thanks for your inputs....
a 500k variable resistor from 6n1P plate to ground, and then the coupling cap to the wiper....
but why not try to just get rid of the 6n1p stage? you do not need that too much gain anyway....or you can parallel 6n1p sections and use that in place of the 6n6p for more voltage gain....
the 13de7 used by Fenris has the 2 triodes with mu of 17.5 and 6 for a theoretical voltage gain of 105, compared to yours with 33 x 22 or 726, so you have way too much gain and for a 5 watter amp, not needed....
but why not try to just get rid of the 6n1p stage? you do not need that too much gain anyway....or you can parallel 6n1p sections and use that in place of the 6n6p for more voltage gain....
the 13de7 used by Fenris has the 2 triodes with mu of 17.5 and 6 for a theoretical voltage gain of 105, compared to yours with 33 x 22 or 726, so you have way too much gain and for a 5 watter amp, not needed....
Last edited:
It would be ideal for the new ortho headphones with a lower ratio transformer, the ones I have take about 2W into 38 ohms. So not as much current as efficient speakers but more voltage.
How much distortion does the mosfet add? Do you have a spice model for it? If they could be found something like a 1980's low amperage hitachi lateral mosfet may improve preformane significantly?
This does drive speakers quite well. The output efficiency is thrillingly low, but it makes up for it by not using a tube in the final stage. Not having a big heater in the output pretty much makes up for the losses in the parafeed arrangement. The MOSFETS as followers pretty much add zero coloration/distortion. Any "tube sound" comes from the distortion specrta of the tube voltage amplification and the properties of the output transformer
Its an promising design because it helps allievate an output transformers weak spot (non-linearity in bass from tubes general output impedance.) It allows the use of smaller ratio output transformers which also give a more linear transformer. Finally the load on the "output" tube is fairly constant and much lower than an output transformer which again increases linearity.
Only drawback is my experience simulating/working with mosfet followers is that they can add significant distortion in certain situations. Would be interested in simulating this topology.
I am thinking of comparing a classic 45SET parafeed output with your mosfet follower with LTspice, but I can't find a model for your mosfet choice. Is there a more common equivalent mosfet?
Thanks again for sharing this, I see potential for gains not just on budget builds (I love the simplicity of a spud amp so nothing against the 13DET) but could consider a semi-budget DHT with a more linear 4P1L serving as the gain input and another 4P1L serving as the output for a semi-budget DHT SET "hybrid" build. Or a classic 7n7-45 parafeed only with your mosfet follower which would allows a very high quality high nickel core OPT for a "no expense build" which might top some expensive 300B builds.
I see your idea as important as the "simple SE" because of allowing higher quality transformers. For example with the right mosfet one could probably follow a 45 with a magnaquest 2.5k:16 instead of the typical 5k:8, thats a big deal IMO.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.