DIY Audio Projects Forum

I have no idea how to answer this question. The amp has an excellent sound. The 6V6 in SE-UL operation is very smooth and true. However, due to the Rp of a pentode, even run in UL mode, the bass -3dB point is right around 40Hz. This is more than sufficient for almost all types of music. The only place where it may be wanting is in some electronic music where the bass is artificially extended in the low end.

If you choose to build, I recommend you use the optimized operation points used in the amp here: http://diyaudioprojects.com/Forum/viewtopic.php?f=9&t=5736

_________________
Matt
It's all about the Glass!
http://www.CascadeTubes.com
Cascade Tubes Blog

Thanks for the explanation. I got it. Well, I don't hear to electronic music at all and even if I go for it, I don't worry because I don't even know if I can hear below 40hz. It's all good for me if it's like this.

Have to agree with Matt 40hz for me is the lower 3db point in my room. My high effciency speakers really don't go any lower anyhow and in classical music (which I listen to) balanced bass is more important than silly blooming bass. Even in jazz groups boomy bass doesn't really exist in live listening. It's all about balance.

_________________
Projects:"retro2308" - chip based headphone amp | ”Calibre 834” - tube phono MM preamp | ”najah” - Raw 180W Tripath Class D power amp | "Icon" - Shuguang CV-181Z preamp

I agree your point.

Hi, previous lurker, first time poster. I'm creating an amp based Suncalc's beautiful 6V6 1.0 design, and I had some questions about AC response, component selection / safety.

6V6 Clipping: If I turn the pot / volume too high, the 6V6 clips. I see two ways to mitigate this.
(1) I can increase my cathode voltage via the series R so that the grid can hit higher amplitues without clipping. However, that has downside of taking my 6V6 out of the datasheet specified -12.5V bias range.
(2) Alternatively, I could obviously reduce input voltage. For this design, that means reducing amplitude via the potentiometer. So either limiting myself to 20 or 30% of the pot's full scale range (cutting into my volume control), or a new smaller pot to reduce load for the 6SN7 triodes to reduce it's gane.
Based on this, I am planning to get a new Pot to reduce 6SN7 gain and reduce amplitude input to the 6V6. However, I wanted to see if there are other design methods I'm missing to keep my amplifiers in linear region and happy. Are there perhaps any tricks with the ultra-linear screen grid tap, or the 6SN7 biases that I am missing?



Low Frequency Response: In the second revision of his design, Suncal shows some really nice low frequency response (viewtopic.php?f=9&t=5736&start=20). Again, I'm sure his real world measurements are far from sims, but my design doesn't get anything even close to that. I'm seeing pretty poor gain up to 200 Hz. This is using spice models from DuncanAmps (http://www.duncanamps.com/spicemodels.html), which could definitely be a source of error, but being new, I am more inclined to think the error is with my design. Can anyone point me to a resource, or suggest some tips, to improve my low frequency gain?



Component Selection: Lastly, based on experience, if anyone has suggestions on how I could improve my component selection please let me know. I chose tubes based off thetubestore reviews, chokes/transformers from this thread, and passives by attempting to way overrate everything.

Here's my design. All components are purchased and on my workbench, but I have not yet powered on. Due to high voltages, I want to understand everything a little bit more fully before I start working through gofar99's Power On guide (viewtopic.php?f=9&t=3676).

Ok. I'm sitting in an airport but let me see if I can give you some points to consider.

First, your schematic is not accurate to either the baseline Lacewood or the V2 version. It looks like you used the driver from V2 and the power stage from the baseline. This is why you don't have too much volume control play before the power stage goes into grid conduction. I strongly recommend the V2 version. It is a better design and takes advantage of the UL mode optimmization work documented on the information page.

Second, I know people like the Duncan amp tools and models, but 30 years as an Electrical Engineer has taught me to be very wary of models and simulation. There are generally wrong much more often than they are right. And once a tube model breaks into grid conduction, the results from the model are useless. Also, the break over is idealized so the harmonic content is all wrong as the drive levels increase.

Take a look at the V2 schematic here http://diyaudioprojects.com/Forum/viewtopic.php?f=9&t=5736&start=13 and let me know if you still have the same problems.

_________________
Matt
It's all about the Glass!
http://www.CascadeTubes.com
Cascade Tubes Blog

Thanks for the feedback.


I will take another look at how I hooked things up; my intention was to follow your 1.0 design. But more than follow X or Y design, I mostly want to make sure I understand how it's all working. This is all new to me. After doing some more poking around, I think my issue as you point out is in poor power stage implementation.


It is good to know they are not accurate and that I still have hope of having some sort of bass come through. And I agree not to use them as a bible. But as someone who didn't know what a grid was a few weeks ago, just trying to use them as a first order tool to gauge understanding.


Both design are saved to my desktop and referenced constantly

Maybe a few more nights poking around then I'll start building it up. Looking forward to turning these tubes on.

Actually I think your problem is mixed design points.

In a two stage amp of this type you need to make sure the driver is matched to the power stage. You have a power stage biased at about 12.5v. Since line level is about 2V peak that means you need very little gain in the driver. The original Lacewood driver stage had about 8v/v. This means that it takes 12.5/8 or about 1.56v peak to drive the power stage into grid conduction.

When you used the mixed driver you did (RL=100k, Rk=1k, bypassed), you ended up with significantly more gain and no volume control play because of too much sensitivity. When you run your simulation, you either need to use one schematic or the other. Mixing the two is a bad idea.

_________________
Matt
It's all about the Glass!
http://www.CascadeTubes.com
Cascade Tubes Blog

Be careful with that. The roll off you referenced is ONLY the INTER-STAGE roll off. This is the portion of the low end response due to the coupling capacitor between the two stages. It is NOT the end to end low frequency response. I never posted the total response curve for this rewire. I will see if I can get that posted this afternoon.

_________________
Matt
It's all about the Glass!
http://www.CascadeTubes.com
Cascade Tubes Blog

Here is the measured frequency response of the Lacewood V2.This is actually pretty typical of pentode UL-mode amplifiers using smaller output transformers. It's down just about -1dB at 100Hz, -3dB at 40Hz, and rolls of smoothly.

Your plot above looked to be about -5dB at 100Hz and about -12dB at 40Hz. Clearly there is something seriously wrong in the model of the power tube, the output transformer, or both. This is one of the reasons I don't put too much stock in models. Just too many unknowns.

Don't be sucked in by the "20 to 20k" crowd. I think you will find this amplifier very detailed and and relatively fast. A good overall listening amplifier.

_________________
Matt
It's all about the Glass!
http://www.CascadeTubes.com
Cascade Tubes Blog