What a nice surprise that you release a new video this soon. Your graphs really do tell a lot. I thought it was just a placebo effect that my voltage follower opamp makes the amp sounded better. And I really appreciate you as well, Mike. Thank you.
Nice little trick I picked up from my Quad preamp, is varying the feedback in an op-amp circuit with a pot. It is also worth noting that solid state amplifiers with FET front ends have high input impedance.
Hi Michael, it's good to see you back again. Thank you for a very interesting and informative video. You have confirmed what I suspected - that the high(ish) impedance of volume controls can affect the H.F roll-off. I didn't realise it was so drastic though. In my homemade L12 amplifier, I use seven 600 ohm resistor attenuators set to -1dB, -2dB, -4dB, -8dB, -16dB and -32dB, switched by reed relays. Opamp buffers between each one ensures a constant impedance and good matching. To me at least it sounds superb. P.S. I love the L12. Thank you for the tip.
Great Sir Michael, I sometimes add buffers, either dual or single, in my power amps but never tone controls (active or passive) because people tend to mess with tone controls, causing sound coloration even though we have good tone controls.❤😊 Love you sir Michael you are a good man.
Having built a few amps and a preamp from scratch, designing impedance match signal circuitry is very important and often overlooked. A modern preamp's main task is to balance input and source impedances, and output/amplifier input impedance. And attenuation of course. The only ways of doing this with potentiometers is to use a buffer either side of the pot. Pots are inherently imperfect in other ways and will never sound as good at 9 o'clock vs. 2 o'clock.
Also a very compelling additional argument for proper gain management. The more that the potentiometer attenuates, the larger the effect. The common "I can only turn it up to 11:00 O'Clock" just means that we have wasted half of the ability to set volume (in trade for HF attenuation). The audio industry has not made life convenient by having digital devices output to a 2V standard and analog devices to .75V. In this case, we even need a buffer with gain control for each input! Buffers can be a great thing if they are truly neutral and have low output impedance. Low impedance also means that the interconnects should have less effects. GREAT video. The jabs at valve technology are in my view pretty unnecessary. Many of us know about cathode followers to dramatically lower output impedance and the ability to get very low noise levels. My point is not in defense of valves, rather that there is no need to muck up a perfectly good solid state video with "half truths" that do not strengthen the otherwise great arguments..
Don't take my comments too seriously, just a bit of fun. It's not a religion, we are all free to think what we do about valves. I will not think of your opinion as wrong or any less of you because your opinion is different to mine. Just enjoy the equipment you have. Still friends???
@@MichaelBeeny I really appreciate your vlog style and content. I just consider the valve jabs as ill placed. I use a Pass transistor preamplifier (low impedance...) connected to single ended pentode valve power amps and my phono corrector is also a Tim DeParavicini valve unit - noise is in spite of the gain, simply not an issue. That is my first hand experience over 2 decades. I measure and listen. Your recent C3850 clone experiences make me want to build one as a headphone amplifier (AKG 701 - around 120 ohms). I will report with measurements and listening experience when I am finished.
Another good and interesting analysis. This is why the input RF shunt caps should always be *before* the volume control. Typically they work best right on the RCA or XLR jacks themselves. With the RF bypass after the volume control ... you effectively end up with a low pass filter, as you explained, that will vary it's roll off with the impedance of the top half of the volume control. Also 50k is too high for the initial level control on an amplifier... try 10k ... see how that goes.
@@peterlarkin762 It's called "impedance bridging" if it needs a name. The general principle is that a low impedance always feeds a higher one. Mike found one of the many reasons why. en.wikipedia.org/wiki/Impedance_bridging
@@peterlarkin762 Yes, that is quite a common practice for DIY'ers and it does make things better. Unfortunately this will not completely resolve the issue when using these boards in their as-shipped unmodified state. When using a 10K volume pot at 50% true setpoint (5K-5K) that is connected directly to the board input, it will still drop by about -2.2dB at 20khz... not "hi-fi" at all. And if you go too low on the volume control resistance it may overload whatever signal source is feeding it or cause other issues. The better solution is to have the signal input connected to the volume control (regardless of its value) without any paralleled capacitance, then through a preamp or buffer amp with a low impedance output, and then connected directly to the amp board inputs. Modifying the amp boards themselves in the manner Mike has described will help a LOT as well. BTW, while placing the RF shunt caps directly at the input jacks on the box is indeed *theoretically* best, other issues can arise when doing that. Some signal sources simply will not tolerate feeding directly into a capacitive loads (even if the capacitance is relatively small). They can become unstable or distorted, and if nothing else they may overshoot by quite a lot. For example, with a mere 100pF paralleled load Mike's favorite op amp (NE5532) can overshoot by up to *10%* on a small signal per the manufacturer's own data sheet..! This is despite the NE5532 having a typical output impedance of merely 0.3 ohms (again per their data sheet). For this reason many (but not all) preamp and source device outputs (especially where op amps are used) have series resistors that appear upon first glance to have no real or obvious purpose. One could suppose that it is for short circuit protection, but that function is already part of the op amp's internal circuitry. Their actual function is to help keep the output stages stable if the load is highly capacitive ("highly" being a relative term here). Series resistors such as this are even present in the outputs of many (if not most) power amps, including the dual TDA7293 boards in Mike's recent videos, and for exactly the same reasons. The only exception to this that I can think of are phono cartridge preamp inputs. They not only tolerate direct capacitive loads, but the cartridge manufacturers usually specify parallel capacitive load values to achieve proper operation.
@@ronschauer839 Yep... getting carried away with RF bypassing on your projects inputs can destabilize some of the less carefully designed driver stages. But fortunately it doesn't take much series resistance to prevent this. The input bypasses with 1k in series then a 47pf capacitor to ground usually do a nice job when feeding a 10k pot which then feeds a 20k or higher input load. There is still a roll off of course, but that should put it up around 50khz where nobody will hear it.
@@Douglas_Blake_579 Without low capacitance in the input section of the power amp (assuming you are still connecting the volume pot directly to it), the roll off may be much more severe than you may imagine. It's not only about the resistive impedance, it's also about the capacitive reactance. 100 ohms is a very typical value for a series resistor at the output of an op amp or DAC stage, by the way. And in the case of the dual7293 boards it is 4.7 ohms with approximately a 1.5uH inductor in parallel with it (to pass lower frequencies unimpeded). However fitting a suitable high input impedance / low output impedance preamp or buffer in between the pot and the power amp input will usually resolve volume control-induced roll off problems nicely, even with the utterly ridiculous 2200pF C1 value originall fitted on those boards. Wait for Mike's next video or two, it will prove this concept I'm sure.
I’d rephrase your description to: “I look at the effect of a standard pot on the sound of a *preamp* when connected directly to the *output*”. That might make it more clear that the output impedance of the prior stage is the key issue. Some folks think the point of this video is “pots are bad”. I don’t read that into at all.
Agree 100%.. pots are just resistive elements..it's how the designer of the circuit, be it internal or external from the amp, can handle the impedances and their matching..! 👍
Agreed. This is one of many videos on this channel presented in such as way as to be potentially very misleading. I'm just glad this guy wasn't my electronics tutor when I started out 50 years ago as my career would have been doomed from the start.
It's a personal preference really. A switch using fixed resisters can be a better tracker, but you do have clicks and fixed jumps in volume. Personally, I like a smooth rotation, without indents. All of these will sound the same, however. Just don't use a digital one, they DO affect the sound, some more than others.
@@MichaelBeeny I have a Lightspeed LDR made here in Oz = I find it very good. I also have an Optivol LDR based pot designed by Greg Ball in Brisbane. Alas he ceased production. It has the funny characteristic of increasing the volume A LOT if you don't power it off first!
Hi Mike. It''s a brave man indeed who decries the use of our thermionic friends in hifi. Hahaha. Do you get much flac for such statements? Of course, I can only agree with you. Apart from a few very high power uses [such as radar] the only credible use of valves in this, the third millenium, is in guitar amplifiers where hiss is [largely] irrelevant and soft clipping is actually a desirable characteristic. Keep 'em comin'. Many thanks. Phil.
Hi Philip. Surprisingly very little. I guess if you have $10/20,000 to spend on 1940s technology that really does look good, why not. Just sad that you can do a better job in ALL respects with an amplifier like the 7293. I guess you could always put it in a gold-plated box, that put the cost up quite abit. Having been in retail HiFi, worked for Quad, Cambridge Audio, Monitor Audio etc to name but a few. What the hell do I know lol
Hi Michael, I was wondering whether your TDA7293 power amp input has a few pico farad capacitor from input to ground ? Then with the volume pot served as a resistor, it will forming an RC filter, thus attentuate high frequency.
The issues here are the reason I use a Krell KRC preamp, it uses a custom 300 resistor volume network and has a discrete class A 15 ohm output. Excellent low volume channel matching is another big bonus of not using a pot.
It's not the fact it's a standard pot that causes the issues; it's the implantation of the attenuator in the circuit. However, fancy the attenuator is it will not make any difference. I cannot comment further as I don't know the circuit in question.
Video does not disappoint as usual...I am always wondering why in the boards I build , when i use the voume control of the pre-amp if I pause a song and dial up the volume the amplifier really starts to show the hiss factor? Is this a impedance issue? Perhaps answers on your upcoming video...i will be staying tuned off course
It could be impedance issues but could also be too much gain from the preamp. Normal listening should be about 10 o'clock, getting very loud at 2-3 o'clock. If it's very loud at 9/10 o'clock you might have too much gain and you might well hear noise at say 4/5 o'clock.
I would NEVER put any kind of transfomer anywhere in any audio circuit. This video is about placing a pot in-between the preamp and the amplifier. It's OK if the preamp has a low output impedance, many don't.
I did look at one a few years ago. I know that was just one example, but It needed so much external parts to make it work well I did not like it. They are used as part of a general preamp common these days. Often full of software and other tone shaping circuits.
Greeting from Canada Mike. I enjoyed your videos on the C3850 clone and scooped one up from Ali. I have no idea if the output of the preamp is Hi or Low impedance and was going to use a potentiometer before feeding it into my power amp. Would you recommend using a buffer amp after the pot then feeding the buffer into the power amp? Have an awesome day and thank you 🙂
The pot should go on the input of the preamp, use a maximum of 50 K, better with a a 25K if you can find one. Dont put it on the input of your power amp, it really messes the HF range of the amplifier. If the gain of the preamp is too high, I have shown how to change it. Always use the minimum gain you can. You don't need anything else; this is a buffer with gain.
@@MichaelBeeny Thank you so much for the quick reply. I will reduce the enthusiasm of the preamp per your video and plumb it right into the power amp. I ordered a Alps 20K and 50K from Ali, one of which I will put right at the front of the preamp. Thank you for your inspirational videos. Ian.
Hi Mike, very interesting experiments. Let me ask you, I got an 80´s Luxman receiver and the dual volume pot became unbalanced, so each channel has a different volume. It´s a bespoke (motorized) pot that I will not find another, so I am considering an Alpine ALPS replacement, losing the motorized option, unfortunately. Well, the question is the original pot is a 150K log and ALPS only offers 100K or 250K. With 250K I will probably have a more "fine tunable" volume control while with the 100K a more aggressive one. But in which one you think I could have the lesser low volume impedance matching issues ? Thanks in advance my friend.
Input impedance, output impedance, damping factor... these are generally too profound for the general consumers. Anyway, nice to demonstrate the effect of a passive attenuation in front of a power amp. I remember a hifi designer ever mentioned he preferred to set his cd63 at low output level. Perhaps that was related to the max-ING of the pot on the preamp stage
Hi Mark, I love building valve kits, built many in the past. As a youth I built all the Mullard valve circuits starting the 3 watt EL84 kit right up the flag ship 20watt EL34 kit which I used for about 5 years. Was one of the best in the day. Things have moved on sound wise; they simply don't stack up these days and cost silly money. However, 3 years ago I did make THIS 10 watter ua-cam.com/video/4WFCfLvLZqs/v-deo.htmlsi=z-hlHp4nGha9iiDT
This problem with pots needs a buffer stage after the pot that has very little input capacitance thus avoiding the low pass filter formed by the pot resistance and amplifier input capacitance. You have to have a volume control in there somewhere.
i am 90% self taught and for as long as i remember noticed this issue but couldnt make sense of it other than guessing it was an impedance mismatch but no idea how to fix it but i always liked to use very low input impedance like 560r to 1k to eliminate mains hum in diy amps since ground loops happen even when all earth references are from the center "tap" of the smoothing caps and found that made very dull sounding amps, lots of pitfalls! took me years to figure out it was my fault for earthing differential input pairs via badly run pcb traces , now i can run 47k input without any hum (unless into headphones) and the sound is full of life again (hitachi 2sj50, 2sk135 opamp style amp)
Hi mate , for once I disagree with your comments. I have some high end active and passive pre amps (NVA power amps , exposure and Quad) with HP and ex BBC test gear . With careful consideration of impedance matching , the passive pre amps perform better with no measurable high frequency roll off . I look forward to your next posts with interest…keep them coming . PS I connect pots in shunt mode with signal passing through a fixed value resistor.
Hi Dave, This video was really a generalization of the topic. Of course, we do have many alternatives ways to the control of volume. Not many people have the knowledge to be able do the things you have suggested or indeed much in the way of test gear you have. My channel is really aimed at the average DIY person. I appreciate your input Dave.
It's never a good idea to feed a pot into the input of most amplifiers unless it already contains a preamp or buffer. I have made follow up videos on this subject already, I go into great detail of different value pots and the effect on frequency responses.
Ok, before even watching the video: There's a reason they used double T (impedance constant) attenuators on audio test equipment. Problem: They go all over the place once the pots inevitably wear down and get scratchy, worse than a simple voltage divider :)
Potentiometers are performance killers in many cases, alas. I once had a very low distortion amplifier, I decided to add a potentiometer when I relocated it to a room where I had no preamp. THD+N jumped 10X! I guess that’s the problem with 0.000X % distortion: The phase of the moon, your mood or the color of your socks can make it worse significantly. 🤣
Thank you for the video. Yes and no. First we like distortion, Nelson Pass proved that with listeners and he designed his amps for those second harmonics listeners like. He did it with fets, I do it with tubes, both give good results. I have this little preamp on order but I have the sense that it will sound 'dry' the same way a class D amp can sound dry. It's the preamp that gives the sound distinct color and we like color, well harmonics. This is why guitar players love tube amps and I personally like that sound along with most other listeners. If you want an amp that gives you out exactly what goes in this is an easy task, any INA (instrument ) opamp will do that and very linearly and accurately, just as accurate and linear as your little amp does probably even more so, it's quite easy. The trouble is musicality is not a product of low distortion, in fact, the human ear really can't hear distortion below about 0.5%, most people can't hear 1%. So these manufacturers bragging about 0.0001% distortion are just making, well - noise! Tubes can be used as preamps and very well. You can design tube outputs to give low output impedance, a well designed cathode follower will give low impedance and a better PSRR. The Aikido circuit explains this well. I won't post a link here but Google "The low impedance Aikido" , he gives a good explanation of this. Personally I use hybrid tube circuits which essentially buffer the tube with a buffer chip or an audio opamp that is capable of driving most amps and even headphones. I've done this with fets also - Google my "Merlin" amp that I did for AudioXpress several years ago. It's an fet design that's flat out to ~250kHz with square waves showing the excellent response.
I have an TDA7498E power amp (which you tested recently ua-cam.com/video/NAW5EFqFYDg/v-deo.html) and simple preamp with one NE5532 (the one with fixed gain and without tone controls - you also tested it in some on your videos ua-cam.com/video/67P18zL1WV8/v-deo.html). Question: should I use potentiometer which already exists on 7498E board? But it it probably in the input... Does it suffer from same impedance mismatch problem? Or should I add other volume potentiometer in front of the preamp?
I wolud have few quaestions before I could agree or not. Like how equal is impedance of meaurement gear in function of frequency. - in case when it decereases significantly, like say there is serial capacitor for DC separation on AC range it's voltage drop on serial resistance of upper part of potentiometer could vary in volume and then it could be measurement issue. Lack of sopranos at lower volume is also obvious due to Fletcher Munson curves. I may say genial solution to it was application of taps and circuitry for loudness correction 20 kHz is not Hi FI standard - it is useless because it is way above Hi Fi standard,. People can hear 20 kHz with special headphones in total silence of shells . . In normal listening it is overlayed by daily noise and by times stronger many harmonics of instruments and other sounds in band of 1kHz - 10 kHz . The same stars people can see at night but stars are not inputting to picture of sky in daytime. Total silence and listening to music at home are conditions like night and day. Very fortunately that in total silence young people can notice higher tones but not everyone has to be capable - it is usefull when we enjoy very silent room at night to hear mosquito ( kidding - it is like 2-3 kHz ) . But at usual listening sound levels we do not hear even 3 mosquitos at 2 kHz We need at least bee at 250 Hz . All we need from sopranos in HI FI is max 10 kHz Fortunately that young people in total silence can notice higher tones but it is not useful for Hi FI music. ( HI FI is standard for conditions in which people are not capable to find difference between natural sound and reproduced by electronic gear)
If the source impedance were zero and the load impedance was infinite the shunt capacitance and series inductance of the connecting cable and potentiometer would be irrelevant. But they aren't. The plate output of a vacuum tube can be 5K to 15K while the input of a solid state amplifier can be as low as 5k to 10k. This is horrible. Now the parameters of what connects them matters a great deal. A cathode follower would be a better idea. However the best idea for all low level outputs is a small class AB amplifier with an output impedance under one ohm. Some manufacturers of solid state amplifiers are using tube's for the input stage. 100k is typical for an input stage of a tube. However there is nothing good that can be said of vacuum tubes in audio circuits. The single most distortion producing device in a tube audio amplifier is probably the output transformer. It produces a type of distortion due to the hysteresis lag in the polarity reversal inertia in the magnetic domains in the core that transistor amplifiers can't duplicate. They really are awful sounding IMO. So a good input stage might be a MOSFET. The most primitive amplifier design is the single ended triode with no feedback. It produces almost no power and is highly unstable. Thermionic emission of the cathode varies with the fourth power of the temperature. For some reason there are people who love this design especially when implemented with the Western Electric 300B vacuum tube. Western Electric is the manufacturing division of AT&T. Because of the demand for these museum piece relics of a century ago I heard they might start manufacturing them again. This design is probably much like the first electronic amplifier Deforest invented.
Interesting stuff about valves and valve amps. It did not occur to me that the iron cored output transformer would have that effect. I only recently came across the hysteric property of silicon steel laminations (not sure if the grain orientation makes a difference to this characteristic). But I have to pull you up on inductance being a factor on the input of an amp. Given that these cables are typically coax they would have to be quite long before the inductance became an issue. Even just a single wire would have to be metres long for inductance of the wire to be big enough to impact the audio frequency range by interacting with other impedances. This is why parasitic inductance it is usually ignored in audio circuits.
@smbd2010 The cores of transformers are thin lamination laminated with an insulating layer between them to reduce eddy current losses. These are induced circulating currents due to the flow of electrons in the windings transverse to the lamina. It generates heat and because of the way it's packaged it's hard for the heat to escape. Still heat loss is usually only 5 to 10 percent of the power transfer. The maximum power theorem tells us that the maximum power is transferred from the source to the load when their impedance are equal. So the primary windings have to match the impedance of the output tube's and the secondary windings must match the impedance of the speakers. That's why there are multiple taps.
@@markfischer3626 Yeah, you are mostly correct. Your understanding of eddy currents needs a little work and the way a transformer works thermally is just plain wrong.. Mate, I have decades of experience in power electronics as a professional EE and really do not need some remedial education in the basics of transformers. I recently designed my first iron core laminated transformer, all the others have been ferrite cores. But I know how it is done and how they work and in based on your reply, I know it far better than you do. The "maximum power transfer theorem" is where people who do not understand what they are talking about go to hide. You have really quite missed the point of the output transformer. The transformer is not there to present two impedances, one to speaker and one to amp. If you must think of this in power transfer theorem terms, the transformer is there to transform the impedance of the speaker to the correct impedance for the amplifier., It is an impedance transforming thing. The impedances of the windings is a whole other thing and a much more nuanced thing. Do you mean the self inductance of each winding? Are you referring to the magnetising current? The leakage inductance? Ultimately, the ideal transformer has windings with no impedance of their own. Now, please do not lecture me like I am a newbie benefiting from your slightly flawed explanations. And know what you are talking about and be correct or expect a response from me. Misinformation presented in authoritative tones is my pet hate as is techno bamboozling by people who do not actually understand anything.
A few months ago, I decided to close my channel, after having a triple bypass. I was OK for a while, but heart failer, and other very major issues meant my life expectancy and quality of life was very limited. Making videos was a VERY low priory, I even started to sell some of my test gear. I now have a few good days in-between hospital trips and feeling very unwell. Thats when I make the odd video. So, my discretion of a largely closed channel but with the odd video now and again when heath allows is quite a good, I thought.
No tone controls! I could not agree more. A decent system and speakers and trust the sound engineer and artist to have done their job properly. Listen to your recordings as they are supposed to be heard. That is my take on this anyway.
Setting aside the dozens of potentiometers that will have been used in the making of your recordings ... the only way you will hear the recordings "as they are supposed to be heard" is if you listen using the exact same amplification and speakers used by the person who did the mastering... any other system will inevitably present the sound with audible differences from the mastering system, even using "a decent system and speakers".
@@smbd2010 You have a choice. For people with an excellent HiFi system that works well on most recordings just leaving it flat then that's great. And it will likely sound great ... but let's not kid ourselves that we are hearing exactly what the mastering engineer heard. If the "HiFi" system is less than stellar then there's more likelihood you will want to use (shock horror) potentiometer controlled tone controls!
@@MrAdopado So, you are expecting to correct for all of the harmonic and inter modulation distortion,, the frequency modulated compression and the resonances etc etc with a treble and bass tone control which is also going to cause all manner of frequency dependent phase shifts which will play havoc with the staging or imaging (vernacular is optional)? Is it really that simple? Gosh! Even on a cheap system, I am still a purist. Tone controls will only add colour. This may matter less for rock or pop music but for orchestral music it is rarely a good thing to attempt 'correction' with tone controls. I am sorry if that is too controversial for you. Mr Beeny points out in many of these videos, the essential characteristics of better systems and as an EE I like that he does that. This video is no different. A buffer after a volume pot to get the impedance back down and independent of the setting is a good sign of a lesser system that cut one too many corners.
@@smbd2010 Where did I say I expected to correct harmonic or inter-modulation distortion? I never specified music types either. My point is simply that the recordings will include the effects of multiple potentiometers before it gets anywhere near your playback system. No matter how pure your playback system you can't avoid this.
What a nice surprise that you release a new video this soon. Your graphs really do tell a lot. I thought it was just a placebo effect that my voltage follower opamp makes the amp sounded better.
And I really appreciate you as well, Mike. Thank you.
Nice to see you back Michael, keep up the good work.
Regards
Dave. 😀
Nice little trick I picked up from my Quad preamp, is varying the feedback in an op-amp circuit with a pot.
It is also worth noting that solid state amplifiers with FET front ends have high input impedance.
Hi Michael, it's good to see you back again. Thank you for a very interesting and informative video. You have confirmed what I suspected - that the high(ish) impedance of volume controls can affect the H.F roll-off. I didn't realise it was so drastic though. In my homemade L12 amplifier, I use seven 600 ohm resistor attenuators set to -1dB, -2dB, -4dB, -8dB, -16dB and -32dB, switched by reed relays. Opamp buffers between each one ensures a constant impedance and good matching. To me at least it sounds superb. P.S. I love the L12. Thank you for the tip.
Great to see you're doing well!! Keep them videos coming.
Hi Mike. Loving to see you and the new content. Keep it up
Great video! Loved the discussion at the end.
Thank you, kind sir.
Great Sir Michael, I sometimes add buffers, either dual or single, in my power amps but never tone controls (active or passive) because people tend to mess with tone controls, causing sound coloration even though we have good tone controls.❤😊
Love you sir Michael you are a good man.
Thank you for your kind words, most appreciated.
Having built a few amps and a preamp from scratch, designing impedance match signal circuitry is very important and often overlooked. A modern preamp's main task is to balance input and source impedances, and output/amplifier input impedance. And attenuation of course.
The only ways of doing this with potentiometers is to use a buffer either side of the pot.
Pots are inherently imperfect in other ways and will never sound as good at 9 o'clock vs. 2 o'clock.
And we appreciate you as well Michael
Thankyou Terence, I appreciate you appreciating me lol.
Also a very compelling additional argument for proper gain management. The more that the potentiometer attenuates, the larger the effect. The common "I can only turn it up to 11:00 O'Clock" just means that we have wasted half of the ability to set volume (in trade for HF attenuation). The audio industry has not made life convenient by having digital devices output to a 2V standard and analog devices to .75V. In this case, we even need a buffer with gain control for each input!
Buffers can be a great thing if they are truly neutral and have low output impedance. Low impedance also means that the interconnects should have less effects.
GREAT video. The jabs at valve technology are in my view pretty unnecessary. Many of us know about cathode followers to dramatically lower output impedance and the ability to get very low noise levels. My point is not in defense of valves, rather that there is no need to muck up a perfectly good solid state video with "half truths" that do not strengthen the otherwise great arguments..
Don't take my comments too seriously, just a bit of fun. It's not a religion, we are all free to think what we do about valves. I will not think of your opinion as wrong or any less of you because your opinion is different to mine. Just enjoy the equipment you have. Still friends???
@@MichaelBeeny I really appreciate your vlog style and content. I just consider the valve jabs as ill placed. I use a Pass transistor preamplifier (low impedance...) connected to single ended pentode valve power amps and my phono corrector is also a Tim DeParavicini valve unit - noise is in spite of the gain, simply not an issue. That is my first hand experience over 2 decades. I measure and listen.
Your recent C3850 clone experiences make me want to build one as a headphone amplifier (AKG 701 - around 120 ohms). I will report with measurements and listening experience when I am finished.
Another good and interesting analysis.
This is why the input RF shunt caps should always be *before* the volume control. Typically they work best right on the RCA or XLR jacks themselves.
With the RF bypass after the volume control ... you effectively end up with a low pass filter, as you explained, that will vary it's roll off with the impedance of the top half of the volume control.
Also 50k is too high for the initial level control on an amplifier... try 10k ... see how that goes.
A lot of diy'er solid state amp designs are 10k or 22k input impedance so 4.7k or 10k potentiometers are definitely recommended.
@@peterlarkin762
It's called "impedance bridging" if it needs a name. The general principle is that a low impedance always feeds a higher one. Mike found one of the many reasons why.
en.wikipedia.org/wiki/Impedance_bridging
@@peterlarkin762
Yes, that is quite a common practice for DIY'ers and it does make things better.
Unfortunately this will not completely resolve the issue when using these boards in their as-shipped unmodified state.
When using a 10K volume pot at 50% true setpoint (5K-5K) that is connected directly to the board input, it will still drop by about -2.2dB at 20khz... not "hi-fi" at all.
And if you go too low on the volume control resistance it may overload whatever signal source is feeding it or cause other issues.
The better solution is to have the signal input connected to the volume control (regardless of its value) without any paralleled capacitance, then through a preamp or buffer amp with a low impedance output, and then connected directly to the amp board inputs.
Modifying the amp boards themselves in the manner Mike has described will help a LOT as well.
BTW, while placing the RF shunt caps directly at the input jacks on the box is indeed *theoretically* best, other issues can arise when doing that.
Some signal sources simply will not tolerate feeding directly into a capacitive loads (even if the capacitance is relatively small).
They can become unstable or distorted, and if nothing else they may overshoot by quite a lot.
For example, with a mere 100pF paralleled load Mike's favorite op amp (NE5532) can overshoot by up to *10%* on a small signal per the manufacturer's own data sheet..!
This is despite the NE5532 having a typical output impedance of merely 0.3 ohms (again per their data sheet).
For this reason many (but not all) preamp and source device outputs (especially where op amps are used) have series resistors that appear upon first glance to have no real or obvious purpose.
One could suppose that it is for short circuit protection, but that function is already part of the op amp's internal circuitry.
Their actual function is to help keep the output stages stable if the load is highly capacitive ("highly" being a relative term here).
Series resistors such as this are even present in the outputs of many (if not most) power amps, including the dual TDA7293 boards in Mike's recent videos, and for exactly the same reasons.
The only exception to this that I can think of are phono cartridge preamp inputs.
They not only tolerate direct capacitive loads, but the cartridge manufacturers usually specify parallel capacitive load values to achieve proper operation.
@@ronschauer839
Yep... getting carried away with RF bypassing on your projects inputs can destabilize some of the less carefully designed driver stages. But fortunately it doesn't take much series resistance to prevent this. The input bypasses with 1k in series then a 47pf capacitor to ground usually do a nice job when feeding a 10k pot which then feeds a 20k or higher input load. There is still a roll off of course, but that should put it up around 50khz where nobody will hear it.
@@Douglas_Blake_579
Without low capacitance in the input section of the power amp (assuming you are still connecting the volume pot directly to it), the roll off may be much more severe than you may imagine.
It's not only about the resistive impedance, it's also about the capacitive reactance.
100 ohms is a very typical value for a series resistor at the output of an op amp or DAC stage, by the way.
And in the case of the dual7293 boards it is 4.7 ohms with approximately a 1.5uH inductor in parallel with it (to pass lower frequencies unimpeded).
However fitting a suitable high input impedance / low output impedance preamp or buffer in between the pot and the power amp input will usually resolve volume control-induced roll off problems nicely, even with the utterly ridiculous 2200pF C1 value originall fitted on those boards.
Wait for Mike's next video or two, it will prove this concept I'm sure.
Really excellent and clear explanation. ❤
Hi Mike. Really enjoy your videos. Thank you for making them ✌️👍
Thank you for watching and commenting, much appreciated.
Mr Beeny complete clear, as always good information to avoid bad builds.
Great video. I learned something today. I think I'm going to do some of this testing for myself.
What happens if i remove the input low pass filter from the amplifier ? Can i get good passive filter ?
I’d rephrase your description to: “I look at the effect of a standard pot on the sound of a *preamp* when connected directly to the *output*”. That might make it more clear that the output impedance of the prior stage is the key issue.
Some folks think the point of this video is “pots are bad”. I don’t read that into at all.
Agree 100%.. pots are just resistive elements..it's how the designer of the circuit, be it internal or external from the amp, can handle the impedances and their matching..! 👍
Agreed. This is one of many videos on this channel presented in such as way as to be potentially very misleading. I'm just glad this guy wasn't my electronics tutor when I started out 50 years ago as my career would have been doomed from the start.
Thank you Michael. Would you care to comment on stepped attenuators and LDR volume controls?
It's a personal preference really. A switch using fixed resisters can be a better tracker, but you do have clicks and fixed jumps in volume. Personally, I like a smooth rotation, without indents. All of these will sound the same, however. Just don't use a digital one, they DO affect the sound, some more than others.
@@MichaelBeeny I have a Lightspeed LDR made here in Oz = I find it very good. I also have an Optivol LDR based pot designed by Greg Ball in Brisbane. Alas he ceased production. It has the funny characteristic of increasing the volume A LOT if you don't power it off first!
Hi Mike.
It''s a brave man indeed who decries the use of our thermionic friends in hifi. Hahaha.
Do you get much flac for such statements?
Of course, I can only agree with you. Apart from a few very high power uses [such as radar] the only credible use of valves in this, the third millenium, is in guitar amplifiers where hiss is [largely] irrelevant and soft clipping is actually a desirable characteristic.
Keep 'em comin'.
Many thanks.
Phil.
Hi Philip. Surprisingly very little. I guess if you have $10/20,000 to spend on 1940s technology that really does look good, why not. Just sad that you can do a better job in ALL respects with an amplifier like the 7293. I guess you could always put it in a gold-plated box, that put the cost up quite abit. Having been in retail HiFi, worked for Quad, Cambridge Audio, Monitor Audio etc to name but a few. What the hell do I know lol
With a back story like that, I am sure you are a veritable fount of hifi anecdotes and knowledge.
Thanks for sharing.
Hi Michael, I was wondering whether your TDA7293 power amp input has a few pico farad capacitor from input to ground ? Then with the volume pot served as a resistor, it will forming an RC filter, thus attentuate high frequency.
Yes, it does, like most amplifiers. That's really one of the points of the video. It's not an issue if fed from a low impedance.
The issues here are the reason I use a Krell KRC preamp, it uses a custom 300 resistor volume network and has a discrete class A 15 ohm output. Excellent low volume channel matching is another big bonus of not using a pot.
It's not the fact it's a standard pot that causes the issues; it's the implantation of the attenuator in the circuit. However, fancy the attenuator is it will not make any difference. I cannot comment further as I don't know the circuit in question.
@@MichaelBeeny Understood.
Video does not disappoint as usual...I am always wondering why in the boards I build , when i use the voume control of the pre-amp if I pause a song and dial up the volume the amplifier really starts to show the hiss factor? Is this a impedance issue?
Perhaps answers on your upcoming video...i will be staying tuned off course
It could be impedance issues but could also be too much gain from the preamp. Normal listening should be about 10 o'clock, getting very loud at 2-3 o'clock. If it's very loud at 9/10 o'clock you might have too much gain and you might well hear noise at say 4/5 o'clock.
What's your opinion on an autoformer,s adjusting the volume from the speaker output or even between the pre and power section?
I would NEVER put any kind of transfomer anywhere in any audio circuit. This video is about placing a pot in-between the preamp and the amplifier. It's OK if the preamp has a low output impedance, many don't.
could you also look critical at volume digital volume control in eg a Wiim streamer?
I did look at one a few years ago. I know that was just one example, but It needed so much external parts to make it work well I did not like it. They are used as part of a general preamp common these days. Often full of software and other tone shaping circuits.
Greeting from Canada Mike. I enjoyed your videos on the C3850 clone and scooped one up from Ali. I have no idea if the output of the preamp is Hi or Low impedance and was going to use a potentiometer before feeding it into my power amp. Would you recommend using a buffer amp after the pot then feeding the buffer into the power amp? Have an awesome day and thank you 🙂
The pot should go on the input of the preamp, use a maximum of 50 K, better with a a 25K if you can find one. Dont put it on the input of your power amp, it really messes the HF range of the amplifier. If the gain of the preamp is too high, I have shown how to change it. Always use the minimum gain you can. You don't need anything else; this is a buffer with gain.
@@MichaelBeeny Thank you so much for the quick reply. I will reduce the enthusiasm of the preamp per your video and plumb it right into the power amp. I ordered a Alps 20K and 50K from Ali, one of which I will put right at the front of the preamp. Thank you for your inspirational videos. Ian.
Hi Mike, very interesting experiments. Let me ask you, I got an 80´s Luxman receiver and the dual volume pot became unbalanced, so each channel has a different volume. It´s a bespoke (motorized) pot that I will not find another, so I am considering an Alpine ALPS replacement, losing the motorized option, unfortunately. Well, the question is the original pot is a 150K log and ALPS only offers 100K or 250K. With 250K I will probably have a more "fine tunable" volume control while with the 100K a more aggressive one. But in which one you think I could have the lesser low volume impedance matching issues ? Thanks in advance my friend.
Love this ,no option to subscribe? Thanks.
Good to have you onboard William, welcome.
Input impedance, output impedance, damping factor... these are generally too profound for the general consumers.
Anyway, nice to demonstrate the effect of a passive attenuation in front of a power amp.
I remember a hifi designer ever mentioned he preferred to set his cd63 at low output level. Perhaps that was related to the max-ING of the pot on the preamp stage
So I guess we're not going to be building a kit Tube / Valve amp then ? That's a shame.
Hi Mark, I love building valve kits, built many in the past. As a youth I built all the Mullard valve circuits starting the 3 watt EL84 kit right up the flag ship 20watt EL34 kit which I used for about 5 years. Was one of the best in the day. Things have moved on sound wise; they simply don't stack up these days and cost silly money.
However, 3 years ago I did make THIS 10 watter
ua-cam.com/video/4WFCfLvLZqs/v-deo.htmlsi=z-hlHp4nGha9iiDT
This problem with pots needs a buffer stage after the pot that has very little input capacitance thus avoiding the low pass filter formed by the pot resistance and amplifier input capacitance.
You have to have a volume control in there somewhere.
i am 90% self taught and for as long as i remember noticed this issue but couldnt make sense of it other than guessing it was an impedance mismatch
but no idea how to fix it
but i always liked to use very low input impedance like 560r to 1k to eliminate mains hum in diy amps since ground loops happen even when all earth references are from the center "tap" of the smoothing caps
and found that made very dull sounding amps, lots of pitfalls!
took me years to figure out it was my fault for earthing differential input pairs via badly run pcb traces , now i can run 47k input without any hum (unless into headphones) and the sound is full of life again (hitachi 2sj50, 2sk135 opamp style amp)
Hi mate , for once I disagree with your comments. I have some high end active and passive pre amps (NVA power amps , exposure and Quad) with HP and ex BBC test gear . With careful consideration of impedance matching , the passive pre amps perform better with no measurable high frequency roll off . I look forward to your next posts with interest…keep them coming .
PS I connect pots in shunt mode with signal passing through a fixed value resistor.
Hi Dave,
This video was really a generalization of the topic. Of course, we do have many alternatives ways to the control of volume. Not many people have the knowledge to be able do the things you have suggested or indeed much in the way of test gear you have. My channel is really aimed at the average DIY person. I appreciate your input Dave.
So i should bypass it?
It's never a good idea to feed a pot into the input of most amplifiers unless it already contains a preamp or buffer. I have made follow up videos on this subject already, I go into great detail of different value pots and the effect on frequency responses.
👍👍 🙂
Thanks again!
Yes and transistors are much easyer on the budgit.
Ok, before even watching the video: There's a reason they used double T (impedance constant) attenuators on audio test equipment. Problem: They go all over the place once the pots inevitably wear down and get scratchy, worse than a simple voltage divider :)
Potentiometers are performance killers in many cases, alas. I once had a very low distortion amplifier, I decided to add a potentiometer when I relocated it to a room where I had no preamp. THD+N jumped 10X! I guess that’s the problem with 0.000X % distortion: The phase of the moon, your mood or the color of your socks can make it worse significantly. 🤣
Yes, the colour of your socks can be a real issue, where distortion is concerned. White socks tend to offer the lowest I find.
@@MichaelBeenyEheh
Thank you for the video.
Yes and no. First we like distortion, Nelson Pass proved that with listeners and he designed his amps for those second harmonics listeners like. He did it with fets, I do it with tubes, both give good results. I have this little preamp on order but I have the sense that it will sound 'dry' the same way a class D amp can sound dry. It's the preamp that gives the sound distinct color and we like color, well harmonics. This is why guitar players love tube amps and I personally like that sound along with most other listeners.
If you want an amp that gives you out exactly what goes in this is an easy task, any INA (instrument ) opamp will do that and very linearly and accurately, just as accurate and linear as your little amp does probably even more so, it's quite easy. The trouble is musicality is not a product of low distortion, in fact, the human ear really can't hear distortion below about 0.5%, most people can't hear 1%. So these manufacturers bragging about 0.0001% distortion are just making, well - noise!
Tubes can be used as preamps and very well. You can design tube outputs to give low output impedance, a well designed cathode follower will give low impedance and a better PSRR. The Aikido circuit explains this well. I won't post a link here but Google "The low impedance Aikido" , he gives a good explanation of this.
Personally I use hybrid tube circuits which essentially buffer the tube with a buffer chip or an audio opamp that is capable of driving most amps and even headphones. I've done this with fets also - Google my "Merlin" amp that I did for AudioXpress several years ago. It's an fet design that's flat out to ~250kHz with square waves showing the excellent response.
I have an TDA7498E power amp (which you tested recently ua-cam.com/video/NAW5EFqFYDg/v-deo.html) and simple preamp with one NE5532 (the one with fixed gain and without tone controls - you also tested it in some on your videos ua-cam.com/video/67P18zL1WV8/v-deo.html).
Question: should I use potentiometer which already exists on 7498E board? But it it probably in the input... Does it suffer from same impedance mismatch problem?
Or should I add other volume potentiometer in front of the preamp?
I wolud have few quaestions before I could agree or not. Like how equal is impedance of meaurement gear in function of frequency. - in case when it decereases significantly, like say there is serial capacitor for DC separation on AC range it's voltage drop on serial resistance of upper part of potentiometer could vary in volume and then it could be measurement issue.
Lack of sopranos at lower volume is also obvious due to Fletcher Munson curves. I may say genial solution to it was application of taps and circuitry for loudness correction
20 kHz is not Hi FI standard - it is useless because it is way above Hi Fi standard,. People can hear 20 kHz with special headphones in total silence of shells . . In normal listening it is overlayed by daily noise and by times stronger many harmonics of instruments and other sounds in band of 1kHz - 10 kHz .
The same stars people can see at night but stars are not inputting to picture of sky in daytime. Total silence and listening to music at home are conditions like night and day. Very fortunately that in total silence young people can notice higher tones but not everyone has to be capable - it is usefull when we enjoy very silent room at night to hear mosquito ( kidding - it is like 2-3 kHz )
. But at usual listening sound levels we do not hear even 3 mosquitos at 2 kHz We need at least bee at 250 Hz . All we need from sopranos in HI FI is max 10 kHz Fortunately that young people in total silence can notice higher tones but it is not useful for Hi FI music. ( HI FI is standard for conditions in which people are not capable to find difference between natural sound and reproduced by electronic gear)
If the source impedance were zero and the load impedance was infinite the shunt capacitance and series inductance of the connecting cable and potentiometer would be irrelevant. But they aren't. The plate output of a vacuum tube can be 5K to 15K while the input of a solid state amplifier can be as low as 5k to 10k. This is horrible. Now the parameters of what connects them matters a great deal. A cathode follower would be a better idea. However the best idea for all low level outputs is a small class AB amplifier with an output impedance under one ohm. Some manufacturers of solid state amplifiers are using tube's for the input stage. 100k is typical for an input stage of a tube. However there is nothing good that can be said of vacuum tubes in audio circuits. The single most distortion producing device in a tube audio amplifier is probably the output transformer. It produces a type of distortion due to the hysteresis lag in the polarity reversal inertia in the magnetic domains in the core that transistor amplifiers can't duplicate. They really are awful sounding IMO. So a good input stage might be a MOSFET.
The most primitive amplifier design is the single ended triode with no feedback. It produces almost no power and is highly unstable. Thermionic emission of the cathode varies with the fourth power of the temperature. For some reason there are people who love this design especially when implemented with the Western Electric 300B vacuum tube. Western Electric is the manufacturing division of AT&T. Because of the demand for these museum piece relics of a century ago I heard they might start manufacturing them again. This design is probably much like the first electronic amplifier Deforest invented.
Interesting stuff about valves and valve amps. It did not occur to me that the iron cored output transformer would have that effect. I only recently came across the hysteric property of silicon steel laminations (not sure if the grain orientation makes a difference to this characteristic). But I have to pull you up on inductance being a factor on the input of an amp. Given that these cables are typically coax they would have to be quite long before the inductance became an issue. Even just a single wire would have to be metres long for inductance of the wire to be big enough to impact the audio frequency range by interacting with other impedances. This is why parasitic inductance it is usually ignored in audio circuits.
@smbd2010 The cores of transformers are thin lamination laminated with an insulating layer between them to reduce eddy current losses. These are induced circulating currents due to the flow of electrons in the windings transverse to the lamina. It generates heat and because of the way it's packaged it's hard for the heat to escape. Still heat loss is usually only 5 to 10 percent of the power transfer. The maximum power theorem tells us that the maximum power is transferred from the source to the load when their impedance are equal. So the primary windings have to match the impedance of the output tube's and the secondary windings must match the impedance of the speakers. That's why there are multiple taps.
@@markfischer3626 Yeah, you are mostly correct. Your understanding of eddy currents needs a little work and the way a transformer works thermally is just plain wrong.. Mate, I have decades of experience in power electronics as a professional EE and really do not need some remedial education in the basics of transformers. I recently designed my first iron core laminated transformer, all the others have been ferrite cores. But I know how it is done and how they work and in based on your reply, I know it far better than you do.
The "maximum power transfer theorem" is where people who do not understand what they are talking about go to hide. You have really quite missed the point of the output transformer. The transformer is not there to present two impedances, one to speaker and one to amp. If you must think of this in power transfer theorem terms, the transformer is there to transform the impedance of the speaker to the correct impedance for the amplifier., It is an impedance transforming thing. The impedances of the windings is a whole other thing and a much more nuanced thing. Do you mean the self inductance of each winding? Are you referring to the magnetising current? The leakage inductance? Ultimately, the ideal transformer has windings with no impedance of their own.
Now, please do not lecture me like I am a newbie benefiting from your slightly flawed explanations. And know what you are talking about and be correct or expect a response from me. Misinformation presented in authoritative tones is my pet hate as is techno bamboozling by people who do not actually understand anything.
What if I told you, you've been saying this UA-cam channel is closed. What's the deal? I'd much rather have it open if you're asking
A few months ago, I decided to close my channel, after having a triple bypass. I was OK for a while, but heart failer, and other very major issues meant my life expectancy and quality of life was very limited. Making videos was a VERY low priory, I even started to sell some of my test gear. I now have a few good days in-between hospital trips and feeling very unwell. Thats when I make the odd video. So, my discretion of a largely closed channel but with the odd video now and again when heath allows is quite a good, I thought.
So... Potentiometer on Amp bad... Potentiometer on input of pre-amp OK.... Got it. Kinda what I thought.
No tone controls! I could not agree more. A decent system and speakers and trust the sound engineer and artist to have done their job properly. Listen to your recordings as they are supposed to be heard. That is my take on this anyway.
Setting aside the dozens of potentiometers that will have been used in the making of your recordings ... the only way you will hear the recordings "as they are supposed to be heard" is if you listen using the exact same amplification and speakers used by the person who did the mastering... any other system will inevitably present the sound with audible differences from the mastering system, even using "a decent system and speakers".
@@MrAdopado And you will correct for that difference with a couple of tone controls?
@@smbd2010 You have a choice. For people with an excellent HiFi system that works well on most recordings just leaving it flat then that's great. And it will likely sound great ... but let's not kid ourselves that we are hearing exactly what the mastering engineer heard. If the "HiFi" system is less than stellar then there's more likelihood you will want to use (shock horror) potentiometer controlled tone controls!
@@MrAdopado So, you are expecting to correct for all of the harmonic and inter modulation distortion,, the frequency modulated compression and the resonances etc etc with a treble and bass tone control which is also going to cause all manner of frequency dependent phase shifts which will play havoc with the staging or imaging (vernacular is optional)? Is it really that simple? Gosh!
Even on a cheap system, I am still a purist. Tone controls will only add colour. This may matter less for rock or pop music but for orchestral music it is rarely a good thing to attempt 'correction' with tone controls. I am sorry if that is too controversial for you.
Mr Beeny points out in many of these videos, the essential characteristics of better systems and as an EE I like that he does that. This video is no different. A buffer after a volume pot to get the impedance back down and independent of the setting is a good sign of a lesser system that cut one too many corners.
@@smbd2010 Where did I say I expected to correct harmonic or inter-modulation distortion? I never specified music types either. My point is simply that the recordings will include the effects of multiple potentiometers before it gets anywhere near your playback system. No matter how pure your playback system you can't avoid this.
Well said. This is why I use a gain cell preamplifier.
I cant hear past 13 k anyway.
Me either ... although that's pretty good for a 72 year old 😄
Yet, somehow, when the roll-off Mike describes happens, you do notice it.