[Two questions, choose your favourite! 
]
The most obvious place to use high quality cables is directly in the signal path, from the input to the grids and, in turn, from the outputs to the successive sections.
But where do you draw the lines?
Using Litz silver cables for the heaters would take some seriously good reasoning.
However, how many of you also use relatively high quality wiring in the plate supplies of your amps?
A true Class A has it's supply pretty much directly connected to the output for the entire cylce. And even a Class AB amp has to draw the power for it's output thru the cables it's supply is wired with. Surely cable quality has an effect on transients, fluidity of power delivery and surge current ability?
I have heard good reviews about a number of Van Den Hul's cables and so was considering them. FDE has mention Deskadel but I'm having a lot of trouble finding anything about them in English.
----------------------------------------------------------------------
Finally, how many of you use true tube based voltage regulators? Morgan Jones suggests using the LM317 for plate regulation due to all the problems with stability that tube regulators have.
This seems a bit odd at first. The tube is effectively in series with the regulator. A lot of tube builders would seem to want to avoid putting solid state electronics so close to the signal.
Futher more, he mentions critically damping the regulator for transient rejection. Perhaps I'm missing something, but isn't this a step futher towards a sterile digital sound?
Morgan describes their role as forcing the supply to more accurately track the music. I take it from this that his critical damping statements relate to not allowing the supply's own transient behavior to become part of the signal; as distortion.
But would this not also lead to surges, that would normally induce transients, stopping dead as they're bypassed?
I'm curious because I've been following his advice and I am going to end up with a signficant number of LM regulators in use. This is my first tube amp and I've never experimented with regulators in this way before, so I'm not sure how this is going to sound. Added to that... they need an equally significant number of heatsinks.
The amp will feature two 12AX7 preamp tubes and an EL34 power tube. Later, another 12AX7 will be added.
Controlling each individually will require 8 LM regulators, 4 of which will need to be designed for high voltage use, which requires the use of 4 high voltage transistors... which also need heat sink mounting!
Is there anywhere online that specialising in heatsinks for multiple T022* packages? I'm getting tired of only finding sites for processor heatsinks. I have to be specific, I do not want to start tapping blank heatsinks myself, so they must come with some form of mounting option already implemented.
]The most obvious place to use high quality cables is directly in the signal path, from the input to the grids and, in turn, from the outputs to the successive sections.
But where do you draw the lines?
Using Litz silver cables for the heaters would take some seriously good reasoning.
However, how many of you also use relatively high quality wiring in the plate supplies of your amps?
A true Class A has it's supply pretty much directly connected to the output for the entire cylce. And even a Class AB amp has to draw the power for it's output thru the cables it's supply is wired with. Surely cable quality has an effect on transients, fluidity of power delivery and surge current ability?
I have heard good reviews about a number of Van Den Hul's cables and so was considering them. FDE has mention Deskadel but I'm having a lot of trouble finding anything about them in English.
----------------------------------------------------------------------
Finally, how many of you use true tube based voltage regulators? Morgan Jones suggests using the LM317 for plate regulation due to all the problems with stability that tube regulators have.
This seems a bit odd at first. The tube is effectively in series with the regulator. A lot of tube builders would seem to want to avoid putting solid state electronics so close to the signal.
Futher more, he mentions critically damping the regulator for transient rejection. Perhaps I'm missing something, but isn't this a step futher towards a sterile digital sound?
Morgan describes their role as forcing the supply to more accurately track the music. I take it from this that his critical damping statements relate to not allowing the supply's own transient behavior to become part of the signal; as distortion.
But would this not also lead to surges, that would normally induce transients, stopping dead as they're bypassed?
I'm curious because I've been following his advice and I am going to end up with a signficant number of LM regulators in use. This is my first tube amp and I've never experimented with regulators in this way before, so I'm not sure how this is going to sound. Added to that... they need an equally significant number of heatsinks.
The amp will feature two 12AX7 preamp tubes and an EL34 power tube. Later, another 12AX7 will be added.
Controlling each individually will require 8 LM regulators, 4 of which will need to be designed for high voltage use, which requires the use of 4 high voltage transistors... which also need heat sink mounting!
Is there anywhere online that specialising in heatsinks for multiple T022* packages? I'm getting tired of only finding sites for processor heatsinks. I have to be specific, I do not want to start tapping blank heatsinks myself, so they must come with some form of mounting option already implemented.
Hi
Regarding regulators I use 2 tube based regulators built with a disimilar triode in my preamp, (I use direct coupling in my amp and therefore use one positive and one negative supply voltage). The last stage is connected directly to the regulator trough a series resistor with bypass caps, (electrolytic + polypropylene) to ground, the amplifier stages are connected in the same manner but fed from the B+ point of the succeeding stage. As I have totally 4 stages in the amp the first input stage of the phono amp is fed trough 4 stages of RC filtering , this is also called graded filtering and used extensively in older equipment.
The reason to use a tube based regulator is because it is not actually trivial to build a solid state regulator with the same performance with regard to high frequency performance and low noise level, I also use a 0B2 gas voltage reference tube as voltage reference as this gives me low noise without a lot of extra components.
The result is that I have an amplifier that is completely hum free and with extremely low background noise level, actually so low that many friends have been doubting that it is really a tube based amplifier.
I dont believe it is necessary to use individual regulators for each stage, (even though some other people dont agree with me!)
Basically I dont think the LM317 based solution is the right way to go, solid state solutions must probably be discrete in order to compete even with simple tube regulators but that is just what I believe.
For Heaters I use a standard 7824 regulator it gives very good results and low noise level
Regards Hans
Regarding regulators I use 2 tube based regulators built with a disimilar triode in my preamp, (I use direct coupling in my amp and therefore use one positive and one negative supply voltage). The last stage is connected directly to the regulator trough a series resistor with bypass caps, (electrolytic + polypropylene) to ground, the amplifier stages are connected in the same manner but fed from the B+ point of the succeeding stage. As I have totally 4 stages in the amp the first input stage of the phono amp is fed trough 4 stages of RC filtering , this is also called graded filtering and used extensively in older equipment.
The reason to use a tube based regulator is because it is not actually trivial to build a solid state regulator with the same performance with regard to high frequency performance and low noise level, I also use a 0B2 gas voltage reference tube as voltage reference as this gives me low noise without a lot of extra components.
The result is that I have an amplifier that is completely hum free and with extremely low background noise level, actually so low that many friends have been doubting that it is really a tube based amplifier.
I dont believe it is necessary to use individual regulators for each stage, (even though some other people dont agree with me!)
Basically I dont think the LM317 based solution is the right way to go, solid state solutions must probably be discrete in order to compete even with simple tube regulators but that is just what I believe.
For Heaters I use a standard 7824 regulator it gives very good results and low noise level
Regards Hans
Hi,
Eeka Chu,
You only need to ask me once...
DESKADEL
In the U.K Audiosynthesis should stock the silver hook-up wire from Deskadel, amongst others.
As far as regulators go, for the HT supply I use valve regulators exclusively. Using sand there makes the entire stage sound like sand to my ears.
Contrary to TubeTVR I use a separate PSU per stage where it matters most, namely in the phono stages. (Guess I'm one of those guys, Hans.)
Other than that I find myself in complete agreement with Hans, in fact I often find we adhere to the same design philosphy, to the point it sometimes gets a little scary...
Another good reason to use valve based VRs is that when wired correctly, nothing can happen to damage the circuit should one decide to go south.
They're absolutely foolproof, you can even pull a valve without any damage at all...
From experience with so called high-end valved gear I can only say that 9/10 it's that SS regulator that's going to need replacing...
Count yourself lucky to find out what they used and to find a replacement.
So far I never, ever had a single component fail in a valved reg for 15 years of daily abuse.
Cheers,
Eeka Chu,
You only need to ask me once...
DESKADEL
In the U.K Audiosynthesis should stock the silver hook-up wire from Deskadel, amongst others.
As far as regulators go, for the HT supply I use valve regulators exclusively. Using sand there makes the entire stage sound like sand to my ears.
Contrary to TubeTVR I use a separate PSU per stage where it matters most, namely in the phono stages. (Guess I'm one of those guys, Hans.)
Other than that I find myself in complete agreement with Hans, in fact I often find we adhere to the same design philosphy, to the point it sometimes gets a little scary...
Another good reason to use valve based VRs is that when wired correctly, nothing can happen to damage the circuit should one decide to go south.
They're absolutely foolproof, you can even pull a valve without any damage at all...
From experience with so called high-end valved gear I can only say that 9/10 it's that SS regulator that's going to need replacing...
Count yourself lucky to find out what they used and to find a replacement.
So far I never, ever had a single component fail in a valved reg for 15 years of daily abuse.
Cheers,
eeka chu said:
Hold the phone. None less resistive than rusty old steel lockwire for *any* connections. Takes a big transformer due to all the resistance inside it though...
<Serious>I have to try steel wire some time. Just to give you guys a run for your money (and what money it is).</Serious>
Tim
Thanks for the help with this everyone!
I've spent some time this morning reading a bit more about valves against solid state and, as with everything in audio it seems, the first two statements I read where entirely contradictory. I'll show you what I mean.
I found this first response on a simple text based forum and have added italics to the parts I found interesting. I quote -
"... The reason I used gas regulators vs solid state, is
the ability of gas regulators to have a longer time
constant over the attack and release cycle of regulation.
What that gives you is a more smooth voltage transition
of regulation, not like the"chopping" effect of solid state.
You can hear this effect in the audio signal.
Solid state has no place here and in my opinion is
nothing more than someone trying to re-invent the
wheel by limiting component space and cost. Also,
gas regulators provide DC filtering. Many people don't
realize that fact."
Looks like valves are clearly in the lead. I finish reading and have a look in Tube Cad.
[From an article by Glassware, February 99, on high voltage regulation]
"From the above description of a regulator, it is obvious that an all tube high voltage regulator could be designed and built, but it would prove very complex and probably fail to meet, at very least, our third design goal: extremely low AC noise. Two problem areas for tubes are relatively weak transconductance and susceptibility to microphonics, both of which limit the amount of noise free operation we can expect from a tube. So with all due apologies to the hard core tube fans, here is an all solid-state design."
click here for the article
I am not trying to say one is better than the other, I'm just trying to show the root of my confusion with this.
I'm not sure, since I'm not well educated enough with regards to regulators, but this choice, between valves and solid state, seems to be similar to the choice of rectifier state.
I am more familiar with this in terms of guitar amplifiers. A significant part of 'The American Sound" is usually the use of valve rectifiers that smooth the response of the amplifier out and create a warmer, rounder, thicker sound. Where as 'The British Sound', typified by Marshall, seems to be much more achievable using an amplifier with a solid state, diode based rectifier. Which creates a much more upfront, defined and active sound.
Would I be right in assuming the regulator choice to be similar?
I would be really interested to hear how many of you often choose tube rectification to go along side tube regulation. Or even more interestingly, how many use a combination of solid state with valve based rectification and regulation.
At the moment, I am aiming for the slightly sharper attack offered by solid state rectification. I am using Morgan Jone's method, again, and implementing Hyperfast diodes in the rectifier. These use a series of resistors across them to balance the voltage division and 10n capacitors to prevent ringing and reverse voltage indifferences.
I will be very interested to see how these sound! I suspect they will most likely be even more upfront due to their faster recovery times.
Sorry, I've kind of side tracked! But again, thanks. And special thanks to FDE for the link to the Deskadel cable!
I've spent some time this morning reading a bit more about valves against solid state and, as with everything in audio it seems
, the first two statements I read where entirely contradictory. I'll show you what I mean.I found this first response on a simple text based forum and have added italics to the parts I found interesting. I quote -
"... The reason I used gas regulators vs solid state, is
the ability of gas regulators to have a longer time
constant over the attack and release cycle of regulation.
What that gives you is a more smooth voltage transition
of regulation, not like the"chopping" effect of solid state.
You can hear this effect in the audio signal.
Solid state has no place here and in my opinion is
nothing more than someone trying to re-invent the
wheel by limiting component space and cost. Also,
gas regulators provide DC filtering. Many people don't
realize that fact."
Looks like valves are clearly in the lead. I finish reading and have a look in Tube Cad.
[From an article by Glassware, February 99, on high voltage regulation]
"From the above description of a regulator, it is obvious that an all tube high voltage regulator could be designed and built, but it would prove very complex and probably fail to meet, at very least, our third design goal: extremely low AC noise. Two problem areas for tubes are relatively weak transconductance and susceptibility to microphonics, both of which limit the amount of noise free operation we can expect from a tube. So with all due apologies to the hard core tube fans, here is an all solid-state design."
click here for the article
I am not trying to say one is better than the other, I'm just trying to show the root of my confusion with this.
I'm not sure, since I'm not well educated enough with regards to regulators, but this choice, between valves and solid state, seems to be similar to the choice of rectifier state.
I am more familiar with this in terms of guitar amplifiers. A significant part of 'The American Sound" is usually the use of valve rectifiers that smooth the response of the amplifier out and create a warmer, rounder, thicker sound. Where as 'The British Sound', typified by Marshall, seems to be much more achievable using an amplifier with a solid state, diode based rectifier. Which creates a much more upfront, defined and active sound.
Would I be right in assuming the regulator choice to be similar?
I would be really interested to hear how many of you often choose tube rectification to go along side tube regulation. Or even more interestingly, how many use a combination of solid state with valve based rectification and regulation.
At the moment, I am aiming for the slightly sharper attack offered by solid state rectification. I am using Morgan Jone's method, again, and implementing Hyperfast diodes in the rectifier. These use a series of resistors across them to balance the voltage division and 10n capacitors to prevent ringing and reverse voltage indifferences.
I will be very interested to see how these sound! I suspect they will most likely be even more upfront due to their faster recovery times.
Sorry, I've kind of side tracked! But again, thanks. And special thanks to FDE for the link to the Deskadel cable!
Hi,
If you so choose go for solid state regulation, maybe you will find that the choice is not so easy as it is on paper, it is a very complex question. But whatever you do stay away from LM317 and other 3 pin regulators, especially if you want something that is faster then a tube based solution, these 3 pins types are very slow and depends entirely on the output capacitor to achieve decent output impedance at high frequencies.
Maybe you should compare reverse recovery time for solid state hex fred diodes to tube rectifiers and the need or lack of need of parallell snubber capacitors.
Regards Hans
If you so choose go for solid state regulation, maybe you will find that the choice is not so easy as it is on paper, it is a very complex question. But whatever you do stay away from LM317 and other 3 pin regulators, especially if you want something that is faster then a tube based solution, these 3 pins types are very slow and depends entirely on the output capacitor to achieve decent output impedance at high frequencies.
Maybe you should compare reverse recovery time for solid state hex fred diodes to tube rectifiers and the need or lack of need of parallell snubber capacitors.
Regards Hans
A contrary view
I love tubes, but I keep them out of my power supply. The role of a regulator in a tube amp is to provide stable DC at a low source impedance so that dynamic demands can be met and that stages are decoupled from one another- to the extent that the regulator does that task, it's a good regulator.
Haven't read the article you linked to, but I fully agree with the exerpt. A good solid-state supply will have a lower source impedance and lower noise than a valve supply, otherwise other industries would be using tubes for high-precision low noise laboratory gear, right? Notwithstanding the bleats of the purists, DC is DC. You can objectively know what supply works best with some simple measurements. IMO, most of the objection to the "sound" of a good SS regulator is made by people who are doing an open-loop listening test with the embarrassing knowledge aforehand that there's a chip or a transistor in there sullying the vacuum glory of their amp.
The real test of a supply is in use- the SS supplies I've been using just plain and simple show LESS coupling and noise, lower source impedance, and better efficiency than any tube-based supply I've tried.
I don't understand the analogy to regulation here. Can you clarify this? A HEXFRED raw supply will show the regulator a lower source Z than a tube rectifier-based raw supply. You might want snubber caps with one rather than the other, sure, but why not complain about the tube rectifier needing a heater supply? The caps are cheaper, don't put out heat, and when you're done, you don't have a rather weedy source resistance in series with the raw supply that the regulator has to work hard to overcome.
I love tubes, but I keep them out of my power supply. The role of a regulator in a tube amp is to provide stable DC at a low source impedance so that dynamic demands can be met and that stages are decoupled from one another- to the extent that the regulator does that task, it's a good regulator.
Haven't read the article you linked to, but I fully agree with the exerpt. A good solid-state supply will have a lower source impedance and lower noise than a valve supply, otherwise other industries would be using tubes for high-precision low noise laboratory gear, right? Notwithstanding the bleats of the purists, DC is DC. You can objectively know what supply works best with some simple measurements. IMO, most of the objection to the "sound" of a good SS regulator is made by people who are doing an open-loop listening test with the embarrassing knowledge aforehand that there's a chip or a transistor in there sullying the vacuum glory of their amp.
The real test of a supply is in use- the SS supplies I've been using just plain and simple show LESS coupling and noise, lower source impedance, and better efficiency than any tube-based supply I've tried.
I don't understand the analogy to regulation here. Can you clarify this? A HEXFRED raw supply will show the regulator a lower source Z than a tube rectifier-based raw supply. You might want snubber caps with one rather than the other, sure, but why not complain about the tube rectifier needing a heater supply? The caps are cheaper, don't put out heat, and when you're done, you don't have a rather weedy source resistance in series with the raw supply that the regulator has to work hard to overcome.
Hi Hans,
When you say faster than a tube based solution, do you mean eletrically faster in operation or faster to design?
In the regulator Morgan details, he uses the recommended output capacitor detailed in the manufacturer's pdf, but then critically damps the output using a resistor in series with it, assuming that it's unlikely the tantalum bead output cap will be enough to remove transients in the real world.
I guess this probably doesn't fully take into account output impedance. I don't remember him mentioning it if it does.
Thanks for the advice. I'm reading about it right now!
When you say faster than a tube based solution, do you mean eletrically faster in operation or faster to design?
In the regulator Morgan details, he uses the recommended output capacitor detailed in the manufacturer's pdf, but then critically damps the output using a resistor in series with it, assuming that it's unlikely the tantalum bead output cap will be enough to remove transients in the real world.
I guess this probably doesn't fully take into account output impedance. I don't remember him mentioning it if it does.
Thanks for the advice. I'm reading about it right now!
Much of the reason one uses that resistor is not for damping, but rather for reliability on power-up/power-down. You sacrifice a little in high frequency output Z, but at the impedance levels of a tube circuit, and in frequency regions that are already knocked down by the general 1/f character of music signals, a few ohms is pretty negligable.
Faster in operation, i.e more comnstant output impedance over wide frequency range and good transient performance. What I mean is that it is quite simple to design a tube based regulator with constant impedance up to 100kHz or so, a LM317 or othjer 3 pin type have increasing output impedance from a few 100Hz. To have a constant output impedance I believe is extremely important in order for the regulator to be transparent and not affect the sonic result. It seems that what Morgan Jones suggest, (to use a series resistor) is trying to achieve just that, it sort of counteracts the idea of using a SS regulator in the first place, (modifying a SS regulator to behave like a simple tube based one!)
Maybe I was a bit cryptic but the fact is that a tube diode doesn't have a spec for reverse recovery as there is non, as in Schottky diodes there are no minority carriers present and therefore the diode switch off instantly when reversed, the speed is only dependant on the parasitic capacitance and inductance of the diode package itself.
It doesn't concern regulation it is only that I think a choice of components or solution should be based on all facts and the fact is that the problems that can exist with SS rectifiers, (related to reverse recovery) doesn't exist with tube rectifiers and vice versa, (heat and high voltage drop in tube ones). I use tube rectifiers and tube based regulation in preamps but have never thought about using it in my power amps, (today I only build OTL's).
Regards Hans
Would you be able to reference me to any particularly good designs for tube based regulation?
Morgan Jones explains the method briefly but spends a lot more time with solid state supplies.
As I understand it at present, shorter and softer recovery times help reduce the noise being produced by rectification. To an extent, power supply noise can be reduce to insignificant amounts using filter circuits and regulation.
Earlier this evening, I was reading a page about tube rectifiers that states that tube rectifiers have a much faster rise time than silicon based diodes. Last time I saw a graph of voltage against time for a tube rectifier, it took somewhere in the order of 5 seconds for the rectifier to recover.
Whereas the first can be somewhat cured with filtration and regulation, I have to think much harder of a way to remove the lag of the tube rectifier. I guess a high value capacitor at it's output would help reduce lag amplitude, but it would also increase it's duration and help the formation of large transients.
I believe this is known as the avalanche ability of silicon diodes.
I know that for the particular amplifier I wish to obtain, a rectifier is most likely going to be too squishy in it's rise time for my preference.
Tube rectifiers seem to lead for their recovery, but their rise times appear to be far from short when HEXFRED's have rise times quoted in nanoseconds.
That aside for a second. I may be mistaken, but earlier I think I read something about FRED diodes having an ability, greater than standard SS diodes, to power harmonics? If I remember rightly this was either due to their avalanche capacities or their recovery. I've since lost the page, as I was reading a number at the time. The diode in question was made by IR, but was apparently simply a newer version of the HEXFRED's they sell. SY, if you know of anything related to this, I'd be very interested to hear about it.
This is proving to be a highly useful thread for me. Thanks again!
Probably not, but where do you get these figures from? A LM317 varies from 0.01ohm to several ohms at 100kHz, only if decoupling capacitors are used the output impedance gets lower but for the regulator itself it is not that good, I also noticed that the curve of output impedance is taken at optimum output current, although I haven't measured it is likely to be worse for lower currents that could be the case in a tube preamp.
BTW please note the transient characteristics of the LM317 with severe overshoot which indicates very low phase margin for the regulator loop.
Regards Hans
Hi,
Certainly...
Yet there's another important advantage to using tubed regulators, especially the series ones using VR tubes; they do provide for a natural softstart of the downstream circuit even when a SS rectification scheme is used.
This also prolongs the lifespan of electrolytic caps considerably as well.
Regarding Zo of tubed regulators, while these probably won't impress a SS designer I do believe they're more than adequate for any audio circuit and I've never measured noise figures that would convince me NOT to use them.
The reason most designers don't use them is mainly expense and ignorance IMO.
The best preamps I've heard so far in 20 years of audio invariably used tube rectification and tubed regulation or tubed rectification with LC staged filtering or a combination of the above.
As Hans stated before, with bigger amplifiers and especially OTLs this kind of recipe is almost impossible to imply except for the input stages because of the high current involved.
Futterman did this for the OTL One but what a monster amp that was...
I was lucky enough to listen to an OTL One once in Frankfurt and I must say this was an earopener in the sense that you never had the impression you listened to an amplifier at all.
Cheers,
Certainly...
Yet there's another important advantage to using tubed regulators, especially the series ones using VR tubes; they do provide for a natural softstart of the downstream circuit even when a SS rectification scheme is used.
This also prolongs the lifespan of electrolytic caps considerably as well.
Regarding Zo of tubed regulators, while these probably won't impress a SS designer I do believe they're more than adequate for any audio circuit and I've never measured noise figures that would convince me NOT to use them.
The reason most designers don't use them is mainly expense and ignorance IMO.
The best preamps I've heard so far in 20 years of audio invariably used tube rectification and tubed regulation or tubed rectification with LC staged filtering or a combination of the above.
As Hans stated before, with bigger amplifiers and especially OTLs this kind of recipe is almost impossible to imply except for the input stages because of the high current involved.
Futterman did this for the OTL One but what a monster amp that was...
I was lucky enough to listen to an OTL One once in Frankfurt and I must say this was an earopener in the sense that you never had the impression you listened to an amplifier at all.
Cheers,
And the output Z at 100 kHz (and in the actual audible band) for a tube regulator run without an output cap...? Not that anyone would do such a silly thing either way, which is why I don't understand the relevance of this. And if your church is Saint-Sanscap, you're not exactly limited to three-pin regs- there are some excellent op-amp/pass transistor circuits out there, like the one sitting in my ST-70.
I recommend a simple type of regulator like this: http://www.bonavolta.ch/hobby/en/audio/reg_sup.htm#Tubes 1, (the first tube based regulator but with a gas voltage reference instead of a Zener diode) There are variants of this one using a pentode as lower tube in order to improve regulation but it is usually not that important in a class A stage as it draws constant current. The same basic circuit in many variants can be seen here: http://members.aol.com/sbench/reg1.html and subsequent pages.
Using a triode as error amplifier gives better bandwidth but worse regulation.
For extreme load regulation there is a voltage regulator type described in the book by Walley and Wallman, "tube amplifiers" using positive voltage fedback but it is probably a bad choice for Audio as it is a bit unstable, (I built one of these as practical example in elecronics class at college)
Where did you see this? The rectifier tube itself is very quick and good for several MHz so it doesn't make sense.
Yes there are many solid state diodes that are faster then the fastest rectifier tubes but that is not that important if you want to rectify 50Hz, what is more important is that most SS diodes are slow when turning off and don't stop leading current directly when reversed and they can therefore cause ringing on relatively high frequencies.
Hex fred or diodes with controlled so called soft recovery behaves better then ordinary rectifier diodes and it is easier to remove the ringing by using a RC snubber over each diode.
In my OTL I used an ordinary rectifier bridge before but I got problems with ringing and it was difficult to remove this by snubbers, later I changed to 8ETH06 diodes from IR and succeded in designing snubbers that removed ringing almost 100%.
In my preamp I use tube rectifiers as I think it works best.
Regards Hans
Hi Hans,
I've had a quick look at the links but I'll save them and spend a lot longer reading them later.
I have read this in a number of texts, however, as it's on hand, I quote from Morgan Jones' Valve Amplifiers -
"Hard vacuum valve rectifiers have only one clear advantage over silicon, but this advantage may be sufficient to make us tolerate their foilbles. Thie rise time (time taken to change from 10% to 90%) of their output voltage when fully loaded ~5s, which greater reduces the inrush current of electrolytic capacitors in comparison with semiconductor rectifiers."
[Included is a piccy of an S shaped curve of % voltage against time. It has a flat section at the origin which accelorates into a linear climb towards the correct voltage. As it nears the correct voltage, the line deaccelorates into another flat section at the correct voltage]
True or not, taking only what Morgan Jones has said in his own book, accepting the disadvantages he lists purely for a tube's softstarting ability is questionable when 80 pences worth of NTC thermistor will do exatly the same thing. Even wiring this thermistor, or a standard resistor, across a relay and having it entirely switched out of the current path during operation would be substantially cheaper than using a tube based rectifier purely to produce a softstart.
Please be note that in my last statements, I am not concerned whether the fundamental logic is correct or not, but only in questioning Morgan's logic of tubes and softstarts.
Would this be virtually full base to peak cycles at 7 MHz or just measurable transmission though?
If valve rectification is capable of 7MHz base to peak cycles, this would indicate a rise time roughly one million times faster than Morgan Jones quoted. A questionable error.
It wouldn't be a problem if the load wasn't dynamic, however, the loading of the supply may change at kHz. As the system amplifies different frequencies at different amplitudes, differing currents must pass through the rectifier, altering it's output voltage. With a longer rise time, it takes a tube based rectifier longer to recover from the changes, smoothing the demands together to a comparably greater degree.
I'm not properly versed with high end audio systems, however, with guitar amplifiers, which I am confident with the sound of, the difference between a SS and tube based rectifier is a very clear smoothing effect. Notes seem to form together much more, rather than standing out as soon as they're stuck. Whether I am electronically correct in saying so or not, this does seem to somewhat damage the equipment's ability to amplify dynamic changes. For instance, subtle vibrato, fast chord progressions or any note tapping tricks seem to become smoothed and less defined.
It's similar to a frequency independant decay in the system's ability to responsed.
Do you not experience any smoothing of the output using tube rectifiers and regulators?
I'm not sure if you set out to do so, but in your explaination you seem to be putting forward the case for SS diodes by discussing your work with their ringing characteristics.
The very fact that tube based power supplies have softstart characteristics, where SS supplies would provide an intolerable inrush surge, suggests their comparably poorer ability to respond to changes. Before ever actually measuring specific values.
It's deduction by definition. No component can not be defined as having excellent an surge rejection ability and, at the same instant, be able to pass subtantial alterations, unaltered, to a greater degree than that component which has a poorer rejection ability, a solid state diode for example. Which I'm sure you're far from suggesting anyway.
Again, as with my statements earlier, I am trying to look at this purely from the logic shown so far. I love tubes, but so far, the only benefit tubes appear to provide in a supply is a softstarting characteristic.
I am entirely forgetting their real world appearances.
And again, I come back to two possibilities. I could either make a tube based supply respond to changes as a SS supply would, with a rapid rise ability, or I can make a SS supply behave with the benefits given by a tube based supply, by placing a resistance between it and any sunstantial capacitances, be they internal or in the external circuit.
To achieve a softstart using a SS circuit, a single thermistor would be suffient. To preserve an ultimately resistance free path, the heated thermistor could even be removed once the capacitances are charged. The softstart may be inherent in tube based supplies, but it's a long way to go to suggest that soldering the two legs of a thermistor is a lot of work, especially when compared with the work required to cut holes for tube sockets and the extra wiring needed for it's subsequent heater.
Futher still, the characteristic that produces the softstart ability of tubes can not be removed once the supply is charged. Unlike the thermistor / risistor which can be in a SS supply; removing the rise time of a tube based supply is much tricker.
At present, there is no way I would ever build an entirely solid state amplifier. But in terms of the amplifier's supply, softstart seems to be the only advantage of valves, and even this is not only very, very, easily replicated, but also becomes a disadvantage once the supply is operational.
Perhaps it would help if you explained what it was about valve based supplies that makes a system seem more pleasant for you to listen to. May I take a guess that it is a reduction in it's harshness?
I am not saying I'm right by any stretch, I'm only arguing the side for solid state circuits to understand the benefits of each. If I did not, this would only be a one word discussion right? If tubes where in question, I would then argue for their side. I don't have some personal hatred of anyone who uses tubes, because I enjoy them myself!
I've had a quick look at the links but I'll save them and spend a lot longer reading them later.
I have read this in a number of texts, however, as it's on hand, I quote from Morgan Jones' Valve Amplifiers -
"Hard vacuum valve rectifiers have only one clear advantage over silicon, but this advantage may be sufficient to make us tolerate their foilbles. Thie rise time (time taken to change from 10% to 90%) of their output voltage when fully loaded ~5s, which greater reduces the inrush current of electrolytic capacitors in comparison with semiconductor rectifiers."
[Included is a piccy of an S shaped curve of % voltage against time. It has a flat section at the origin which accelorates into a linear climb towards the correct voltage. As it nears the correct voltage, the line deaccelorates into another flat section at the correct voltage]
True or not, taking only what Morgan Jones has said in his own book, accepting the disadvantages he lists purely for a tube's softstarting ability is questionable when 80 pences worth of NTC thermistor will do exatly the same thing. Even wiring this thermistor, or a standard resistor, across a relay and having it entirely switched out of the current path during operation would be substantially cheaper than using a tube based rectifier purely to produce a softstart.
Please be note that in my last statements, I am not concerned whether the fundamental logic is correct or not, but only in questioning Morgan's logic of tubes and softstarts.
Would this be virtually full base to peak cycles at 7 MHz or just measurable transmission though?
If valve rectification is capable of 7MHz base to peak cycles, this would indicate a rise time roughly one million times faster than Morgan Jones quoted. A questionable error.
It wouldn't be a problem if the load wasn't dynamic, however, the loading of the supply may change at kHz. As the system amplifies different frequencies at different amplitudes, differing currents must pass through the rectifier, altering it's output voltage. With a longer rise time, it takes a tube based rectifier longer to recover from the changes, smoothing the demands together to a comparably greater degree.
I'm not properly versed with high end audio systems, however, with guitar amplifiers, which I am confident with the sound of, the difference between a SS and tube based rectifier is a very clear smoothing effect. Notes seem to form together much more, rather than standing out as soon as they're stuck. Whether I am electronically correct in saying so or not, this does seem to somewhat damage the equipment's ability to amplify dynamic changes. For instance, subtle vibrato, fast chord progressions or any note tapping tricks seem to become smoothed and less defined.
It's similar to a frequency independant decay in the system's ability to responsed.
Do you not experience any smoothing of the output using tube rectifiers and regulators?
I'm not sure if you set out to do so, but in your explaination you seem to be putting forward the case for SS diodes by discussing your work with their ringing characteristics.
The very fact that tube based power supplies have softstart characteristics, where SS supplies would provide an intolerable inrush surge, suggests their comparably poorer ability to respond to changes. Before ever actually measuring specific values.
It's deduction by definition. No component can not be defined as having excellent an surge rejection ability and, at the same instant, be able to pass subtantial alterations, unaltered, to a greater degree than that component which has a poorer rejection ability, a solid state diode for example. Which I'm sure you're far from suggesting anyway.
Again, as with my statements earlier, I am trying to look at this purely from the logic shown so far. I love tubes, but so far, the only benefit tubes appear to provide in a supply is a softstarting characteristic.
I am entirely forgetting their real world appearances.
And again, I come back to two possibilities. I could either make a tube based supply respond to changes as a SS supply would, with a rapid rise ability, or I can make a SS supply behave with the benefits given by a tube based supply, by placing a resistance between it and any sunstantial capacitances, be they internal or in the external circuit.
To achieve a softstart using a SS circuit, a single thermistor would be suffient. To preserve an ultimately resistance free path, the heated thermistor could even be removed once the capacitances are charged. The softstart may be inherent in tube based supplies, but it's a long way to go to suggest that soldering the two legs of a thermistor is a lot of work, especially when compared with the work required to cut holes for tube sockets and the extra wiring needed for it's subsequent heater.
Futher still, the characteristic that produces the softstart ability of tubes can not be removed once the supply is charged. Unlike the thermistor / risistor which can be in a SS supply; removing the rise time of a tube based supply is much tricker.
At present, there is no way I would ever build an entirely solid state amplifier. But in terms of the amplifier's supply, softstart seems to be the only advantage of valves, and even this is not only very, very, easily replicated, but also becomes a disadvantage once the supply is operational.
Perhaps it would help if you explained what it was about valve based supplies that makes a system seem more pleasant for you to listen to. May I take a guess that it is a reduction in it's harshness?
I am not saying I'm right by any stretch, I'm only arguing the side for solid state circuits to understand the benefits of each. If I did not, this would only be a one word discussion right? If tubes where in question, I would then argue for their side. I don't have some personal hatred of anyone who uses tubes, because I enjoy them myself!
I think what Morgan means is that at switch-on it takes 5 secs to start the current flowing - the 7Mhz (if that is true) is in fully heated condition.
The advantage only happens at switch-on, so if you leave you amp on all the time, only once in a lifetime.
I am not a tube expert, but I can imagine that the softstart is more important for the other tubes, giving them also a chance to warm up before the HT comes on line. The caps probably don't have a problem.
Jan Didden
The advantage only happens at switch-on, so if you leave you amp on all the time, only once in a lifetime.
I am not a tube expert, but I can imagine that the softstart is more important for the other tubes, giving them also a chance to warm up before the HT comes on line. The caps probably don't have a problem.
Jan Didden
I'm most definately not an expert, but it occurs to me that this discussion, in omitting the effects of chokes and filter capacitors, is overstating the "dynamic contrasts" that a rectifier will be exposed to - especially if the circuit is class A. This may be more valid for a C-R-C input with minimal filtration supplying a class B circuit.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.