Yes, phase as well. Essentially, the goal is to make the signal channel’s response, both frequency and phase, flatter down to the generally accepted 20Hz lower limit of the ear. 20Hz is simply the generally accepted engineering practice for specification of an audio signal channel’s lower frequency limit. It is not a must meet target parameter, except, maybe, for commercial audio equipment.
That’s not surprising. Many factors are involved in the perception of the deepest bass. Such as, your individual low frequency hearing sensitivity to deep bass, the deep bass capability of your speakers, the deep bass response and modes of your listening room, the deep bass content of the music you listen to, etc.
Yes, it means that it‘s acceptably close enough as far as the amplitude response at 32Hz is concerned, for the non-critical application that is in-home audio playback. Especially, when you consider the above factors making the heard response non-flat.
2Hz is just a rule-of-thumb target. A quasi-arbitrary parameter to reference. There’s nothing absolutely necessary about it. Any more than it’s necessary that all systems have response flat up to 20KHz. For many listeners, signal channel response flat only to, say, 17KHz will be subjectively indistinguishable compared with one that’s flat to 20KHz. Each individual listening circumstance features it’s own unique requirements. 20Hz-20KHz type general parameter targets are for the purpose of setting uniform engineering for the signal channel.
Apparently, phase is much more difficult to distinguish than frequency. Despite some people's concerns about absolute phase, there is at least one site where you can listen to the same song with different phase. Turns out almost nobody can consistently identify the in phase version.
Phase is not something that recording studios pay much attention to either. The songs on a particular album might not all be in the same phase and even tracks on a particular song are unlikely to be in the same phase. Two different tube guitar amps might have different numbers of stages so the output of one might be inverted phase and the other not. The engineers just put a mic in front of each of them and record them the same way.
For sure. I've often recommended that people use a simple online tone generator to determine their actual range of hearing before they get overly concerned about response at the frequency extremes.
As we age we typically lose much of our ability to hear high frequencies and, as I understand it, the decline typically starts when we are still relatively young. That's probably even more the case today, with so many younger people listening with earbuds at high volume.
Of course, many who are into audio are older yet they will insist that they have golden ears and can still hear the full 20Hz to 20KHz range, or even beyond. I'm skeptical. If true, that's wonderful, but they are extreme outliers for sure. Really, there isn't much musical information at the frequency extremes anyway.
I'm 71 and, last I checked, I'm only good to about 12k and even then the perceived amplitude of frequencies at the extremes is much lower than the mids. That's true regardless of the range, of course. That's also why I'm not so concerned about -3db points of elements that create a low pass filter. On the low end I'm good to about 40Hz, which is handy because I play bass and I don't listen to much music that has anything lower than what's produced by a bass.
I don't have the equipment needed to measure frequency response. So I must plead guilty to relying on empirical methods (my ears / perception) though I do like to try to understand what's happening in a technical sense, even if I can't hear a difference. Hence, all my questions.
I concur with those views. Regarding those who believe they have golden ears, assuming they do hear things which casual listeners don’t notice, I‘ve concluded that it’s because they have honed their listening observational skills, rather than their having greater hearing acuity. I feel that phenomena is a fault in many subjective listening tests.
I know that I’m greatly more aware of, more observant of characteristics and aspects of playback sound than I was years ago, even though I also have no doubt that my acuity has gotten worse as I age.
I know that I’m greatly more aware of, more observant of characteristics and aspects of playback sound than I was years ago, even though I also have no doubt that my acuity has gotten worse as I age.
I think that You have? The same device thart You are using to post this forum? Use some software for audio measurements? Even if it is a mobile phone i am pretty sure that it is some software for simple audio measurements of transfer and phase...
You dont need a mic, just cables RCA to 3.5mm banana jack.
Google for some free apps for mobile and computer too.
for instance fuzzmeasure is software for audio measurements for mac osx
You have plenty for windows...
The goal is to keep the phase flat to one digit drift at the edges of tha BW.
with -3db at edge points, the phase will be probably around +-20 deg
with -0.25db, the phase will be probably around +-one digit deg
and it can be done by simple calculation of simple and cheep elements as R (pot) and C.
(But in praxis we usually faced with expensive C, R and POTs used with wrong values... 🤢 claiming that the sound is excellent...)
Fof that simulations with some spice software is helping a lot, to visualize and check calculations.
.
Another thing thatshoud be considered is relation with internal resistance of the source and pot (with the terminating resistors)
Internal resistance of the source connected ti the pot create dumping and compression of the signal. this is about the low driving force... That will lead to the loss of the transparency. IF the tube is the source as DAC with the tube output stage WITHOUT buffer, or RIAA preaplifier again without buffer, the output resistance of these sources will be probably wery high in the range of Kilo ohms without tha tube buffer.
So this will create the "compression potentiometer"
.
Example:
After calculating Dynamic capacitance to determine MAX value of the pot for low levels RC high pass filter, next step is to determine MAX value of source internal resistance as generator.
lets say the tube for preamp is ECC82 and pot value at the input is 22Kohm
Use this link https://sound-au.com/articles/l-pad-calc.htm and Calculate Attenuation Of Existing Network
So for minimum damping of again -0,25db as 22K load we need source to be less than 600 ohms of onternal impedance.
solid state outputs are fit into this BUT tube as the sources NOT if they dont have some tube buffer to lower output resistance.
(If it is not 2A3 used as last stage with something about 600ohms outpur resistance...)
(1/S is the rough output resistance of tube cathode follower as buffer)
.
for 6K of source input resistancde the compression will be expected as more than 2db 🙁
(NOW calculate output resistance of the riaa preamplifier with ECC83 in end stage, in anode follower, probably more than 6K...)
in the pictures 22000 ohms is POT value, Rser of 600ohms is SOURCE output resistance, Rpar is the terminating resistor Rg at the input of tube stage (220K to 2.2meg in case that pot failed to keep the grig grounded...)
Internal resistance of the source connected ti the pot create dumping and compression of the signal. this is about the low driving force... That will lead to the loss of the transparency. IF the tube is the source as DAC with the tube output stage WITHOUT buffer, or RIAA preaplifier again without buffer, the output resistance of these sources will be probably wery high in the range of Kilo ohms without tha tube buffer.
So this will create the "compression potentiometer"
.
Example:
After calculating Dynamic capacitance to determine MAX value of the pot for low levels RC high pass filter, next step is to determine MAX value of source internal resistance as generator.
lets say the tube for preamp is ECC82 and pot value at the input is 22Kohm
Use this link https://sound-au.com/articles/l-pad-calc.htm and Calculate Attenuation Of Existing Network
So for minimum damping of again -0,25db as 22K load we need source to be less than 600 ohms of onternal impedance.
solid state outputs are fit into this BUT tube as the sources NOT if they dont have some tube buffer to lower output resistance.
(If it is not 2A3 used as last stage with something about 600ohms outpur resistance...)
(1/S is the rough output resistance of tube cathode follower as buffer)
.
for 6K of source input resistancde the compression will be expected as more than 2db 🙁
(NOW calculate output resistance of the riaa preamplifier with ECC83 in end stage, in anode follower, probably more than 6K...)
in the pictures 22000 ohms is POT value, Rser of 600ohms is SOURCE output resistance, Rpar is the terminating resistor Rg at the input of tube stage (220K to 2.2meg in case that pot failed to keep the grig grounded...)
Last edited:
You need a sound card and Arta software or REW
Try to use and measure then you will have better ieas
Yes REW is good and free. win and osx versions. Does not need even special sound card just use with lap top with in and out...
Very good covered with how-to use.
Very good covered with how-to use.
Well I got this "color" pre-amp built, wow it sounds fantastic! It adds some nice warmth to any amp that is a little too harsh, SS or tube. I ended up adding a couple of small 100K "trimmer" pots next to each input RCA jack to be able to dial in the amount of "color" it adds from a fixed output device like a CD deck and also allows the balance to be set. Basically am setting those trimmers to reduce the 2VRMS signal from my CD deck about 15-20% seems to be the sweet spot. Great job Matt on this design!
https://www.youtube.com/playlist?list=PLtEhh3UpOsiBkQPULTwRi9-OvvaIx1swq
https://www.youtube.com/playlist?list=PLtEhh3UpOsiBkQPULTwRi9-OvvaIx1swq
Great job, Stephe! Glad you're one of those who openly enjoys "warmth" and color without being apologetic about it. If those damn subjectivist audio magazines can simply just admit they enjoy distortion or coloration - I know I do - without brushing it off with flowery pornographic prose about what they hear to justify their preference, we wouldn't have this endless and fruitless debates about accuracy vs "musicality," subjective vs objective, etc... ad nauseam. It would've certainly saved a lot of ink and paper or trees! The worst and most laughable is besides the denial they go on to tell you how much they value "neutral" or tell you one equipment is on the hot or cold side of neutral!
I noticed in Matt's site that he added a correction in the comment after his hand drawn schematic where he has 5.8V at the cathode when it should be 7V. Did your unit come out the same value at 7v at the cathode?
I noticed in Matt's site that he added a correction in the comment after his hand drawn schematic where he has 5.8V at the cathode when it should be 7V. Did your unit come out the same value at 7v at the cathode?
Glad to hear that it turned out well. I used the same combination - a pair of mono pots on the input and a stereo pot on the output - on my Boogie Factor 1626 preamp. The combination allows for maximum versatility. Use whatever combination of settings sounds best in a particular system. As a bonus you get the added ability to adjust the L-R balance. If you find that attenuation on the input is not needed, the pots have no negative effect.
For whatever it's worth, LTspice seems to think that for this circuit working into a 10k ohm load, at most positions on the volume control, the output will be down -2dB or so at 20Hz.
However, with the volume control turned all the way up, the output at 20Hz will be down -5dB. But the vol ctrl won't be used that way, for sure. The gain of the circuit is 8.4x at 1kHz, so in most systems the vol ctrl will always be turned down low.
With the vol ctrl adjusted for 1V RMS in/1V RMS out into a 10k ohm load, THD is predicted to be 1%.
With 200mV RMS in/200mV RMS out, THD is predicted to be 0.21%.
I know it's just a simulation, but it was quick and easy enough to do, and it does suggest there is likely to be some bass attenuation.
I used Adrian Immler's Philips/ECG 12AU7 model, but the results were very similar using other 12AU7/ECC82 models. I think the above is pretty close to reality.
However, with the volume control turned all the way up, the output at 20Hz will be down -5dB. But the vol ctrl won't be used that way, for sure. The gain of the circuit is 8.4x at 1kHz, so in most systems the vol ctrl will always be turned down low.
With the vol ctrl adjusted for 1V RMS in/1V RMS out into a 10k ohm load, THD is predicted to be 1%.
With 200mV RMS in/200mV RMS out, THD is predicted to be 0.21%.
I know it's just a simulation, but it was quick and easy enough to do, and it does suggest there is likely to be some bass attenuation.
I used Adrian Immler's Philips/ECG 12AU7 model, but the results were very similar using other 12AU7/ECC82 models. I think the above is pretty close to reality.
A 6.8uf cap resolved the low frequency roll off issues. Start at 4:00 to see the measured frequency response with a 6.8uf output cap. I tested this at various volume control settings and I didn't see any change either on the low or high end on the build amp.
I used a 100k pot between the 2 stages. The 2nd stage drives the tone stack in to the mosfet amp. Sounds fantastic.
Yes. With the 1µf output capacitor the -3dB design point is ≈15Hz. This puts the response ≈2.2 dB down at 20Hz. The test data on the project page over at CascadeTubes shows the measured response at 20Hz to be -2.1dB. So the prototype behaves exactly as designed.
And yes, driving a 10kΩ load, at full volume, the -3dB frequency rises to ≈29Hz. But, as you said, this would represent an end to end gain of approximately 20dB. It's unlikely that the preamp would ver be run in such a condition. I have included previously graphs showing such response changes with volume setting for various values of coupling capacitors.
So the question at hand has to do with what is considered a reasonable amount of low end rolloff at 20Hz. Typically that's -3dB for audio reproduction systems. How much is allotted to a piece of equipment such as this may be open to debate. However, in virtually all music, there is precious little content below about 40Hz. The exception to that typical design is in home theater where sound effects can actually go much lower. However special equipment is required to reproduce such frequencies in most rooms. This preamp was design for musical reproduction so this restriction doesn't apply in this case.
Stephe substituted a 6.8µf cap in her build which lowers the -3dB frequency to ≈2.2Hz which results in very minimal loss at 20Hz. This change renders the discussion moot. I generally avoid this corner case (minimum loss at 20Hz) because of uncertainties with the load. If the load is driven into a conduction condition, larger values of coupling capacitance can lead to blocking distortion. In my best Engineering judgement the trade off for a couple dB of response at 20Hz is not warranted. Of course, others may have other opinions.
Personally I tend to avoid computer simulations except under very limited conditions. There are far too many compromises in most SPICE packages for me to ever trust their results. In my experience, basic circuit analysis almost always provides more accurate results with far less effort. That last is simply my opinion on the matter. An opinion I shared with the late Bob Pease for exactly the same reasons he always gave.
Hi
Perhaps it will better to put volume controle between stages.
You just need a capaciter after the volume pot before the cathodyne and you can choose the value doesn't matter 10k to 500k
yan
Perhaps it will better to put volume controle between stages.
You just need a capaciter after the volume pot before the cathodyne and you can choose the value doesn't matter 10k to 500k
yan
So long as you don't over-load the input,
an output volume control will improve your signal to noise ratio.
an output volume control will improve your signal to noise ratio.
That is something I plan to experiment with on a future project! I did discover having an adjustment on the input and the output gives the ability to adjust how much harmonic distortion is added to the signal.