Love to have your input but noting that I've gotten under your skin, I won't expect i
The circuit below shows a (simplified) equivalent circuit of a real transformer.
Rc is the series resistance of the primary.
Ll is the leakage inductance of the primary.
Lp is the inductance of the primary (the ideal transformer has infinite inductance).
Rp represents the losses in the core.
Cd is the capacitance of the primary winding.
The secondary will also have series resistance, inductance, leakage inductance and all these parameters can be measured with respect to the secondary windings as well.
With this definition of terms out of the way:
Large transformer often sound better. The larger cores do not saturate nearly as easily during the peak current demands. Jocko and I have measured this and noticed a subjective correlation between transformer size and sonics. Even for low current circuits like preamps, large transformers tend to sound better. A small resistor in series with the secondary windings to limit peak current often improves the sound of smaller transformers, which tends to confirm the saturation theory. This can be seen looking at the transformer's secondary voltage waveform with a scope.
Paul McGowan discusses the topic of large transformers in a Stereophile interview. The subject is on the first page of:
http://www.stereophile.com/showarchives.cgi?222
Toroidal transformers can sound fine but they do not tolerate DC and low frequency garbage that tends to make them saturate and mechanically buzz. It is not that hard to ADS couple the primary windings with back to back filter caps shunted by diodes to limit reverse bias voltage on the capacitors to a diode drop. Jeff Rowland was doing this 10 years ago as I recall. Toroidal transformers are also designed for high flux density for greatest efficiency which makes them more sensitive to the previously mentioned. I know of one designer that has his custom transformers made with lower flux densities.
Different size and type transformers have different leakage inductance for the secondary windings. The combination of this inductance with rectifier capacitance can form a high frequency resonant circuit creating RFI noise that your amplifier may not like. The use of 1000uF caps with no high frequency bypassed can make this problem even worse since this resonant circuit may be at a frequency above the point were the 1000uF cap becomes inductive again. Paralleling the secondary winding with a capacitor will form a lower frequency resonance with a high Q that can helpful or harmful dependency on the resonant frequency. It is possible that this cap may also provide increased diode switching noise since diodes are discharging the stored charge in the cap as they turn on (and off as well, from the reverse recovery currents) This low impedance cap provides a much lower source impedance that the leakage inductance and winding resistance of the transformers secondary windings and will allow higher diode currents. Winding techniques and core shapes can also provide different degrees of coupling of RFI noise on the AC line to your amplifier/preamplifier circuits and form resonant circuits that ring when excited by RFI from the AC line or diode bridge.
In the course of designing an amplifier with a 60 Hz oscillator to create a clean 120V power source for low level circuits, I found something interesting. The various step up transformers used on the output of the amp sounded different. This was even though the transformer was outside the amp's feedback loop for and did not effect the stability of the amp. The best sounding transformer in this set up was an output transformer from a 200 W Conrad Johnson tube amplifier! The higher bandwidth and better core material may be a factor since the charging current for the diode bridge and capacitor supplies being powered, contain frequencies much higher than the 60 Hz of the AC line voltage.
I don't suppose it surprising that different transformers sound different in light of these and other factors. There are circuits that can minimize these problems whose design should be optimized for the particular transformer used.
It's excitation current not "excitable current"
http://pemclab.cn.nctu.edu.tw/W3elemac/W3slide/ch2.xformer/tsld022.htm
The circuit below shows a (simplified) equivalent circuit of a real transformer.
Rc is the series resistance of the primary.
Ll is the leakage inductance of the primary.
Lp is the inductance of the primary (the ideal transformer has infinite inductance).
Rp represents the losses in the core.
Cd is the capacitance of the primary winding.
The secondary will also have series resistance, inductance, leakage inductance and all these parameters can be measured with respect to the secondary windings as well.
With this definition of terms out of the way:
Large transformer often sound better. The larger cores do not saturate nearly as easily during the peak current demands. Jocko and I have measured this and noticed a subjective correlation between transformer size and sonics. Even for low current circuits like preamps, large transformers tend to sound better. A small resistor in series with the secondary windings to limit peak current often improves the sound of smaller transformers, which tends to confirm the saturation theory. This can be seen looking at the transformer's secondary voltage waveform with a scope.
Paul McGowan discusses the topic of large transformers in a Stereophile interview. The subject is on the first page of:
http://www.stereophile.com/showarchives.cgi?222
Toroidal transformers can sound fine but they do not tolerate DC and low frequency garbage that tends to make them saturate and mechanically buzz. It is not that hard to ADS couple the primary windings with back to back filter caps shunted by diodes to limit reverse bias voltage on the capacitors to a diode drop. Jeff Rowland was doing this 10 years ago as I recall. Toroidal transformers are also designed for high flux density for greatest efficiency which makes them more sensitive to the previously mentioned. I know of one designer that has his custom transformers made with lower flux densities.
Different size and type transformers have different leakage inductance for the secondary windings. The combination of this inductance with rectifier capacitance can form a high frequency resonant circuit creating RFI noise that your amplifier may not like. The use of 1000uF caps with no high frequency bypassed can make this problem even worse since this resonant circuit may be at a frequency above the point were the 1000uF cap becomes inductive again. Paralleling the secondary winding with a capacitor will form a lower frequency resonance with a high Q that can helpful or harmful dependency on the resonant frequency. It is possible that this cap may also provide increased diode switching noise since diodes are discharging the stored charge in the cap as they turn on (and off as well, from the reverse recovery currents) This low impedance cap provides a much lower source impedance that the leakage inductance and winding resistance of the transformers secondary windings and will allow higher diode currents. Winding techniques and core shapes can also provide different degrees of coupling of RFI noise on the AC line to your amplifier/preamplifier circuits and form resonant circuits that ring when excited by RFI from the AC line or diode bridge.
In the course of designing an amplifier with a 60 Hz oscillator to create a clean 120V power source for low level circuits, I found something interesting. The various step up transformers used on the output of the amp sounded different. This was even though the transformer was outside the amp's feedback loop for and did not effect the stability of the amp. The best sounding transformer in this set up was an output transformer from a 200 W Conrad Johnson tube amplifier! The higher bandwidth and better core material may be a factor since the charging current for the diode bridge and capacitor supplies being powered, contain frequencies much higher than the 60 Hz of the AC line voltage.
I don't suppose it surprising that different transformers sound different in light of these and other factors. There are circuits that can minimize these problems whose design should be optimized for the particular transformer used.
It's excitation current not "excitable current"
http://pemclab.cn.nctu.edu.tw/W3elemac/W3slide/ch2.xformer/tsld022.htm
Attachments
Re: Re: What we need most in an Audio Power Transformer
Koinichiwa,
I would think so.
I have not had the chance to extensively test such setups due to 240V AC being standard here and 220V/230V in Europe.
When I had Mains transformers custom made for my large PA Amp's in east germany in the 1980's (on LL cores) we found that with around 20 - 30% more turns the transformers ran with much less vibration and heat and seemed more efficient. Given the cost of doing that in copper in those days it was a major piece of budget even in these Amp's.
For people outside the 100...120V Mains voltage area the requirement would remain for dual transformers, using dual transformers with 115V+115V primaries would allow a simple change of voltage between continents AND it would allow also the balancing out of parasitic leakage capacitances to the chassis.
Using then dual winding full wave rectification is simply a question of specifying the correct higher secondary Voltage. As an example use 35V+35V to give appx 18V+18V AC in the "low flux" connection, just right for gainclones.
Sayonara
Koinichiwa,
Peter Daniel said:
I would think so.
I have not had the chance to extensively test such setups due to 240V AC being standard here and 220V/230V in Europe.
When I had Mains transformers custom made for my large PA Amp's in east germany in the 1980's (on LL cores) we found that with around 20 - 30% more turns the transformers ran with much less vibration and heat and seemed more efficient. Given the cost of doing that in copper in those days it was a major piece of budget even in these Amp's.
For people outside the 100...120V Mains voltage area the requirement would remain for dual transformers, using dual transformers with 115V+115V primaries would allow a simple change of voltage between continents AND it would allow also the balancing out of parasitic leakage capacitances to the chassis.
Using then dual winding full wave rectification is simply a question of specifying the correct higher secondary Voltage. As an example use 35V+35V to give appx 18V+18V AC in the "low flux" connection, just right for gainclones.
Sayonara
Fred Dieckmann
Ditto! Welcome aboard Fred. Okay, it seems like you and Kuei are in agreement on the lower flux core density as being the primary quality to look for in a good sounding transformer. I will restate the order of importance as I understand it so Kuei, feel free to correct me:
1. lower flux core density
2. DC blocking ability
3. low leakage (measured by reading the Aq)
Furthermore, these attributes, in the order presented, will be relevant in evaluating all types of transformers and are universally accepted as the way to a good power supply.
If we are all in agreement, we can begin assessing a weight of importance to each of these attributes and we can begin exploring economic ways to correct their deficiencies. I am quite ethusiastic about what will be discovered in this thread because of the eclectic nature of its readers and the multiple real world experiences of our members. I will let a few days go by for you guys to think about other factors that you may have left out. One that comes immediately to my mind is the wire gauge used to construct the transformer. This may be obvious, but its the kind of thing that shouldn't be assumed as a constant. I'm sure there are other things that we haven't considered and I expect a few days of rumination will bring them out.
Ditto! Welcome aboard Fred. Okay, it seems like you and Kuei are in agreement on the lower flux core density as being the primary quality to look for in a good sounding transformer. I will restate the order of importance as I understand it so Kuei, feel free to correct me:
1. lower flux core density
2. DC blocking ability
3. low leakage (measured by reading the Aq)
Furthermore, these attributes, in the order presented, will be relevant in evaluating all types of transformers and are universally accepted as the way to a good power supply.
If we are all in agreement, we can begin assessing a weight of importance to each of these attributes and we can begin exploring economic ways to correct their deficiencies. I am quite ethusiastic about what will be discovered in this thread because of the eclectic nature of its readers and the multiple real world experiences of our members. I will let a few days go by for you guys to think about other factors that you may have left out. One that comes immediately to my mind is the wire gauge used to construct the transformer. This may be obvious, but its the kind of thing that shouldn't be assumed as a constant. I'm sure there are other things that we haven't considered and I expect a few days of rumination will bring them out.
Koinichiwa,
Sure. I merely pointed out that doing it "right" (I am investigating this currently in detail WRT Valve Amplifier transformers) is quite costly and stops you from being able to use "commodity" Transformers.
Yes, you get a better transformer, but at a very significant cost increase. I think most people rather will add a handfull of components worth a few bucks and double or quadruple the number of used "commodity" transformers as this will still be cheaper than have the theoretically near ideal part made.
I think having a simple universal "conditioning" PCB (or at least a suitable circuit) available which can take the needed diodes and capacitors to block DC, the suggested series RC snubber on the primary to reduce and damp ringing, the suggested common mode LC filtering (good standin for not having electrostatic screens) would be of quite good value to a lot of DIY'ers.
Then perhaps comparative testing between different parts in critical positions (especially CM Chokes and small filter Cap's to go with these) would be helpful too.
Then one can also look at the rectification issue and try the various options using bridges, full wave rectification, seperate sections or centertapped and so on.
The result would perhaps be a "best practice" FAQ for handling the way power gets into our gear. This way a lot of expensive power conditioners can become reduntant, simply because gear is engineered to work well from the start.
Sayonara
Sure. I merely pointed out that doing it "right" (I am investigating this currently in detail WRT Valve Amplifier transformers) is quite costly and stops you from being able to use "commodity" Transformers.
Yes, you get a better transformer, but at a very significant cost increase. I think most people rather will add a handfull of components worth a few bucks and double or quadruple the number of used "commodity" transformers as this will still be cheaper than have the theoretically near ideal part made.
I think having a simple universal "conditioning" PCB (or at least a suitable circuit) available which can take the needed diodes and capacitors to block DC, the suggested series RC snubber on the primary to reduce and damp ringing, the suggested common mode LC filtering (good standin for not having electrostatic screens) would be of quite good value to a lot of DIY'ers.
Then perhaps comparative testing between different parts in critical positions (especially CM Chokes and small filter Cap's to go with these) would be helpful too.
Then one can also look at the rectification issue and try the various options using bridges, full wave rectification, seperate sections or centertapped and so on.
The result would perhaps be a "best practice" FAQ for handling the way power gets into our gear. This way a lot of expensive power conditioners can become reduntant, simply because gear is engineered to work well from the start.
Sayonara
gentle Puck, take this transformed scalp*
DC blocking ability is not a function of the transformer but is achieved but a capacitor circuit in series with the primary winding
"3. low leakage (measured by reading the Aq)"
This is not a feature on which to base merit as larger transformers can have larger excitation current due to lower primary inductance. Low flux density though, can be a good thing and leads to less distortion and mechanical hum and buzz on noisy AC lines.
I have used the split primary 220 volt connection run on 120 volts for several high bias amps and I think it is worth doing. The transformers have less mechanical hum and buzz in this mode. There no reason not to try using two transformers with both the primary windings in series and secondary windings in series. Make sure the secondaries are wired with the correct polarity by measuring the open circuit voltage before applying a load to the secondaries. The wrong polarity will not hurt the transformers but will give a very low voltage. This will double the VA rating of the single transformer with the trade of twice the winding resistance. This can be potentially beneficial by lessening the peak diode current in the rectifiers. I have seen small resistors added to the secondaries in even supplies for power amps. These resistors are used in the output stage power packs
sold at: http://www.schuro.de/preisl-ntke.htm These very good designs for diode bridge/rectifier cap circuits other than the fact one should add series resistors to the snubbers caps. I recommend looking at the http://www.ub-elektronik.de/download/ntke.pdf document.
A lot of these techniques can be use to get the desired voltage from surplus transformers. Some transformers split the primary into 100 and 20 volt windings to use with 100 volt or 120 volt mains differing connections of these windings can give an option of several secondary voltages. I have even used the two 100 V windings in series and added the 20 windings in series with the secondaries. You must observe the polarities though and connect series windings for highest voltage. I often drive one of the windings with a signal generator to get the turns ratio I want between primary and secondary you can adjust to output level of the generator to get 11.5 on the primary and multiply the measurement secondary voltage by 10 to determine the voltage when the primary is driven by AC. This lets you use lower voltages that the signal generator can produce and lets you measure things without using lethal voltages.
MAINS VOLTAGES CAN KILL. USE THE GREATEST OF CAUTION AND DO NOT WORK WITH THESE VOLTAGE IF YOU DO NOT FEEL YOU ARE QUALIFIED.
*And, gentle Puck, take this transformed scalp
From off the head of this Athenian swain;
That, he awaking when the other do,
May all to Athens back again repair
And think no more of this night's accidents
But as the fierce vexation of a dream.
A Midsummer Night's Dream- William Shakespeare
DC blocking ability is not a function of the transformer but is achieved but a capacitor circuit in series with the primary winding
"3. low leakage (measured by reading the Aq)"
This is not a feature on which to base merit as larger transformers can have larger excitation current due to lower primary inductance. Low flux density though, can be a good thing and leads to less distortion and mechanical hum and buzz on noisy AC lines.
I have used the split primary 220 volt connection run on 120 volts for several high bias amps and I think it is worth doing. The transformers have less mechanical hum and buzz in this mode. There no reason not to try using two transformers with both the primary windings in series and secondary windings in series. Make sure the secondaries are wired with the correct polarity by measuring the open circuit voltage before applying a load to the secondaries. The wrong polarity will not hurt the transformers but will give a very low voltage. This will double the VA rating of the single transformer with the trade of twice the winding resistance. This can be potentially beneficial by lessening the peak diode current in the rectifiers. I have seen small resistors added to the secondaries in even supplies for power amps. These resistors are used in the output stage power packs
sold at: http://www.schuro.de/preisl-ntke.htm These very good designs for diode bridge/rectifier cap circuits other than the fact one should add series resistors to the snubbers caps. I recommend looking at the http://www.ub-elektronik.de/download/ntke.pdf document.
A lot of these techniques can be use to get the desired voltage from surplus transformers. Some transformers split the primary into 100 and 20 volt windings to use with 100 volt or 120 volt mains differing connections of these windings can give an option of several secondary voltages. I have even used the two 100 V windings in series and added the 20 windings in series with the secondaries. You must observe the polarities though and connect series windings for highest voltage. I often drive one of the windings with a signal generator to get the turns ratio I want between primary and secondary you can adjust to output level of the generator to get 11.5 on the primary and multiply the measurement secondary voltage by 10 to determine the voltage when the primary is driven by AC. This lets you use lower voltages that the signal generator can produce and lets you measure things without using lethal voltages.
MAINS VOLTAGES CAN KILL. USE THE GREATEST OF CAUTION AND DO NOT WORK WITH THESE VOLTAGE IF YOU DO NOT FEEL YOU ARE QUALIFIED.
*And, gentle Puck, take this transformed scalp
From off the head of this Athenian swain;
That, he awaking when the other do,
May all to Athens back again repair
And think no more of this night's accidents
But as the fierce vexation of a dream.
A Midsummer Night's Dream- William Shakespeare
The main disadvantage of running primaries in series with 120V is lowering rated power of the transformer. It has nothing to do with a core but with the gauge of secondary wire. The ga. is calculated to deliver current in accordance with expected voltage on secondaries, so running secondaries at half voltage requires increasing the gauge of the wire to achieve the same power rating, yet the gauge doesn't change and effectively one can expect lower power because secondaries have hard time with delivering enough current.
I got that info from Plitron technical dept. when I was investigating that, when building Aleph X. So it is advisable for that setup to either get higher than required power of the transformer or as for bigger gauge on secondaries.
I got that info from Plitron technical dept. when I was investigating that, when building Aleph X. So it is advisable for that setup to either get higher than required power of the transformer or as for bigger gauge on secondaries.
Fred Dieckmann
Well, yes but it can still be valid if we use the value as a percentage. For example; to judge two identically rated transformers that have different readings in Aq. If we agree that excitable current can effect the sound of a transformer, then we must allow for the possiblility that less Aq will sound better than more. I don't know this to be true, it very well may be vice versa or a non issue but I think that it deserves exploration. Lets say one 1kVA toroid has Aq=200mA and another has Aq=50, why couldn't these be compared and evaluated?
I don't know that this is true. There may be some transformer manufacturers that build the cap into their transformer winding. The point is that it is an issue with transfromer performance so it was included. Why haven't you commented on the wire gauge used in the construction of the winding? If we are going to use an amp that uses a total bias of 20A, what is the optimum thickness of its winding wire to get the best sound?
X-formerfiles.
Hi,
While windings do have capacitance I don't know of any xformer with a built-in cap.
I would not make too big an issue out of it though.
Normally and unless you specify otherwise, the manufacturer uses a chart to define AWG with respect to VA.
Other factors he'll keep in mind are enamel voltage insulation for HV xformers, temperature and humidity insulation.
Obviously for a 20A rating you're not going to pick a wire thickness that can just about carry that amperage, if you want the xformer to run relatively cool you pick the next AWG size.
There are many more factors coming into play such as core quality, mechanical stability and so on.
To me, a good xformer is cool running, doesn't exhibit any mechaincal noise, is able to respond quickly to current demand and is basically self-regulating.
A decent powerxformer manufacturer is a major key to success, be that commercially or personnally.
While I would agree that a good transformer is expensive, if however sound quality matters than it should matter to you too.
From experience, I never heard of an underrated xformer sounding better than an overrated one, unless perhaps they were of greatly differing quality.
Lets never, ever forget that an amp is nothing more than a modulated PSU. (I'm simplifying for simplicity's sake).
Strange as it may seem I often find that tubeamp PSU xformers are far better made than the xformers used in solid state amps, probably because the tubeamp xformers are often made the old school way.
Also, keep in mind that good core material for an OPT is NOT what you want for a powerxformer, quite the contrary and this is the problem of most toroids...they're bandwidth lets all the PSU crap through, in both directions.
So what's needed is a very poor bandwidth xformer, in fact if it wouldn't let through more than a couple of hundred Hz we'd be fine.
Cheers,
Hi,
While windings do have capacitance I don't know of any xformer with a built-in cap.
I would not make too big an issue out of it though.
Normally and unless you specify otherwise, the manufacturer uses a chart to define AWG with respect to VA.
Other factors he'll keep in mind are enamel voltage insulation for HV xformers, temperature and humidity insulation.
Obviously for a 20A rating you're not going to pick a wire thickness that can just about carry that amperage, if you want the xformer to run relatively cool you pick the next AWG size.
There are many more factors coming into play such as core quality, mechanical stability and so on.
To me, a good xformer is cool running, doesn't exhibit any mechaincal noise, is able to respond quickly to current demand and is basically self-regulating.
A decent powerxformer manufacturer is a major key to success, be that commercially or personnally.
While I would agree that a good transformer is expensive, if however sound quality matters than it should matter to you too.
From experience, I never heard of an underrated xformer sounding better than an overrated one, unless perhaps they were of greatly differing quality.
Lets never, ever forget that an amp is nothing more than a modulated PSU. (I'm simplifying for simplicity's sake).
Strange as it may seem I often find that tubeamp PSU xformers are far better made than the xformers used in solid state amps, probably because the tubeamp xformers are often made the old school way.
Also, keep in mind that good core material for an OPT is NOT what you want for a powerxformer, quite the contrary and this is the problem of most toroids...they're bandwidth lets all the PSU crap through, in both directions.
So what's needed is a very poor bandwidth xformer, in fact if it wouldn't let through more than a couple of hundred Hz we'd be fine.
Cheers,
"So what's need is a very poor bandwidth xformer, in fact if it wouldn't let through more than a couple of hundred Hz we'd be fine."
I don't think I agree. I believe the frequencies of the charging current waveform are higher than this. Limiting the amount of EMI trough the transformer would be nice but the high leakage or high frequency core losses or leakage inductance to achieve this can bring a whole set of new issues to deal with. On my power amp oscillator set up to generate clean AC, the tube output transformer sounded better than the transformers design for 60 Hz.
"problem of most toroids...they're bandwidth lets all the PSU crap through, in both directions."
This is a function over the core geomentry which have lower leakage inductances. It is not a function of the core material.
IE transformers have greater bandwitdhs than you would suppose. Alot of what come through from the line is RF coupling
though interwind inf capacitance. Spilt bobbins minimize this capacitance but have higher leakage inductance than transformers with the secondary windings wound over the primary windings.
"that excitable current can effect the sound of a transformer"
Once again the term is "excitation current" and it is a function of the primary inductance and core losses. This current in the primary also changes with load current and the primary currents due to changing load currents with signal are much higher than the unloaded excitation current. Two transformers that have different excitation currents have different designs and several other parameters that are also different. It really doesn't make sense to talk about the excitation current in isolation from the other design parameters, since changing it, changes several other design parameters at the same time.
"The main disadvantage of running primaries in series with 120 V is lowering rated power of the transformer. It has nothing to do with a core but with the gauge of secondary wire. The gauge. is calculated to deliver current in accordance with expected voltage on secondaries"
Losses in transformers are also related to core losses. The gauge of the secondaries (and primaries) determines the power loss due to IxIxR or so called I squared R losses. These power losses are a function of current not voltage. Don't forget we are talking quality not quantity here. Some additional transformer losses are not a big deal and we are most often are not running a transformer at maximum capacity. Transformer designers are trying to design for the maximum efficiency for a given core size and getting them to design for low flux density and the resulting advantages is a foreign concept to them. My designer friend had a major battle trying to get transformers with low flux density designed since the most important goal is high efficiency. Just as in class A amps efficiency is not the most important factor. What is a disadvantage for efficiency maybe an advantage for sonics, just like in amplifiers.
I don't think I agree. I believe the frequencies of the charging current waveform are higher than this. Limiting the amount of EMI trough the transformer would be nice but the high leakage or high frequency core losses or leakage inductance to achieve this can bring a whole set of new issues to deal with. On my power amp oscillator set up to generate clean AC, the tube output transformer sounded better than the transformers design for 60 Hz.
"problem of most toroids...they're bandwidth lets all the PSU crap through, in both directions."
This is a function over the core geomentry which have lower leakage inductances. It is not a function of the core material.
IE transformers have greater bandwitdhs than you would suppose. Alot of what come through from the line is RF coupling
though interwind inf capacitance. Spilt bobbins minimize this capacitance but have higher leakage inductance than transformers with the secondary windings wound over the primary windings.
"that excitable current can effect the sound of a transformer"
Once again the term is "excitation current" and it is a function of the primary inductance and core losses. This current in the primary also changes with load current and the primary currents due to changing load currents with signal are much higher than the unloaded excitation current. Two transformers that have different excitation currents have different designs and several other parameters that are also different. It really doesn't make sense to talk about the excitation current in isolation from the other design parameters, since changing it, changes several other design parameters at the same time.
"The main disadvantage of running primaries in series with 120 V is lowering rated power of the transformer. It has nothing to do with a core but with the gauge of secondary wire. The gauge. is calculated to deliver current in accordance with expected voltage on secondaries"
Losses in transformers are also related to core losses. The gauge of the secondaries (and primaries) determines the power loss due to IxIxR or so called I squared R losses. These power losses are a function of current not voltage. Don't forget we are talking quality not quantity here. Some additional transformer losses are not a big deal and we are most often are not running a transformer at maximum capacity. Transformer designers are trying to design for the maximum efficiency for a given core size and getting them to design for low flux density and the resulting advantages is a foreign concept to them. My designer friend had a major battle trying to get transformers with low flux density designed since the most important goal is high efficiency. Just as in class A amps efficiency is not the most important factor. What is a disadvantage for efficiency maybe an advantage for sonics, just like in amplifiers.
Re: gentle Puck, take this transformed scalp*
Koinichiwa,
Which would have to be sopping big to allow a Class AB Amp (which for sll intents and purposes virutally ALL solid state Amplifiers are, including many supposed Class A Amplifiers and for example all commercial Pass Amplifiers) to function withou being a hinderance to performance. Using the added diode trick makes the whole thing bearable in terms of component cost of course.
Making alternatively a transformer that inherently, by virtue of an airgap and a larger core can just handle some offset still strikes me inherently as better, because the DC blocker circuit still does only addresses the waveform asymmetry incompletely. I do not dispute the effect of the DC blocker (on the contrary, I think with Torroids it is NOT optional), yet there are more complete ways.
They cary however very substantial cost penalties, which in pure custom made, small scale production transformers (for Valve Amp's) nearly disappear, but which would break many a solid state camels back.
When I talk about low leakage I am refering to two seperate, but related phenomenae.
Leakage type one is best measured by taking a current reading (wideband) between mains earth and the common of the secondary voltage of the whole supply. This is common/differential mode leakage into the PSU caused by transformer imbalances.
When selecting from my large collection of wallwarts one to run the coils of the relais in my passive preamp (did I mention that I have a bad case of compulive disorder) I tested a large number of wall warts for this and found interestingly the lowest leakage with one originally driving a cheapernet repeater. The mobile phone chargers I had hoped to be very low (girlies with wet fingers touching the exposed metal of the plug) in fact where abysimal.
Leakage type two is actually caused by parasitic capacitances between primay and chassis/case/ground. It can be measured again against mains earth this time from the chassis, making sure of course that chassis and secondary PSU common are isolated.
A good trick is to take two identical transformers in an arrangement that is mechanically and electrically identical, excepting of course the phasing. Using opposite phasing on the primarie(s) with symmetrical mechanics and secondary supply wiring will usually reduce the leakage currnets by 20-40db.
Of course, making mains transformers by design better would likely improve things by 60db, but at a cost that is not small. Even medical grade isolation transformers are in my view not sufficiently free of leakage to be really good and ideal as a single unit for audio.
Or wind 20-30% more turns.
Very valid and important words. For the uninitiated DIY'er a really good quality "off the shelf" audio transformer would be good.
Of course, making modern style "power packs" incorporating active rectifiers directly off the mains (not phase cut) and switching DC/DC converters would likely be a better choice, as all major issues are adressed. At the moment the cost is still higher than a basic torroid wound in China with cheap labour, sadly.
Anyway, mains powered supplies are such a hassle that for solid state I'd be tempted to be an envoironmental capitalist, imperialist pig and use Batteries, rachargable or not.
The simple bulk needed to get 450V/0.2A to drive my SE Valve Monoblocks and the 250V/0.12A to drive my phonostage, plus heaters has convinced me to keep improving AC supplies. Why all you hardcore "DIY" guys put up with as crappy mains supplies as you do is beyound me. I was on batteries for Solid State in the mid 1990's....
Six pieces of 6AH/12V Selaed Lead Acid batteries power a perfectly decent 50W/8Ohm Amp with a good Class A area for at leass 12 Hours without recharge and are not much larger than big torroids.
Sayonara
Koinichiwa,
Fred Dieckmann said:
Which would have to be sopping big to allow a Class AB Amp (which for sll intents and purposes virutally ALL solid state Amplifiers are, including many supposed Class A Amplifiers and for example all commercial Pass Amplifiers) to function withou being a hinderance to performance. Using the added diode trick makes the whole thing bearable in terms of component cost of course.
Making alternatively a transformer that inherently, by virtue of an airgap and a larger core can just handle some offset still strikes me inherently as better, because the DC blocker circuit still does only addresses the waveform asymmetry incompletely. I do not dispute the effect of the DC blocker (on the contrary, I think with Torroids it is NOT optional), yet there are more complete ways.
They cary however very substantial cost penalties, which in pure custom made, small scale production transformers (for Valve Amp's) nearly disappear, but which would break many a solid state camels back.
Fred Dieckmann said:
When I talk about low leakage I am refering to two seperate, but related phenomenae.
Leakage type one is best measured by taking a current reading (wideband) between mains earth and the common of the secondary voltage of the whole supply. This is common/differential mode leakage into the PSU caused by transformer imbalances.
When selecting from my large collection of wallwarts one to run the coils of the relais in my passive preamp (did I mention that I have a bad case of compulive disorder) I tested a large number of wall warts for this and found interestingly the lowest leakage with one originally driving a cheapernet repeater. The mobile phone chargers I had hoped to be very low (girlies with wet fingers touching the exposed metal of the plug) in fact where abysimal.
Leakage type two is actually caused by parasitic capacitances between primay and chassis/case/ground. It can be measured again against mains earth this time from the chassis, making sure of course that chassis and secondary PSU common are isolated.
A good trick is to take two identical transformers in an arrangement that is mechanically and electrically identical, excepting of course the phasing. Using opposite phasing on the primarie(s) with symmetrical mechanics and secondary supply wiring will usually reduce the leakage currnets by 20-40db.
Of course, making mains transformers by design better would likely improve things by 60db, but at a cost that is not small. Even medical grade isolation transformers are in my view not sufficiently free of leakage to be really good and ideal as a single unit for audio.
Fred Dieckmann said:
Or wind 20-30% more turns.
Fred Dieckmann said:
Very valid and important words. For the uninitiated DIY'er a really good quality "off the shelf" audio transformer would be good.
Of course, making modern style "power packs" incorporating active rectifiers directly off the mains (not phase cut) and switching DC/DC converters would likely be a better choice, as all major issues are adressed. At the moment the cost is still higher than a basic torroid wound in China with cheap labour, sadly.
Anyway, mains powered supplies are such a hassle that for solid state I'd be tempted to be an envoironmental capitalist, imperialist pig and use Batteries, rachargable or not.
The simple bulk needed to get 450V/0.2A to drive my SE Valve Monoblocks and the 250V/0.12A to drive my phonostage, plus heaters has convinced me to keep improving AC supplies. Why all you hardcore "DIY" guys put up with as crappy mains supplies as you do is beyound me. I was on batteries for Solid State in the mid 1990's....
Six pieces of 6AH/12V Selaed Lead Acid batteries power a perfectly decent 50W/8Ohm Amp with a good Class A area for at leass 12 Hours without recharge and are not much larger than big torroids.
Sayonara
The X-former Files.
Hi,
Sure, I'd like to keep it low though...
Absolutely.
And this opposes our own design goal, we want max. Q for a given price.
I have a hard time believing that...Not only does it fly in the face of common engineering believe, it also contradicts everything else you correctly claim.
Doesn't make any sense to me. Not that I know anything about it ...
Designing for Class A operation is any designer's wet dream...it is sooo easy.
It relaxes most parameters except for costs.
Cheers,
Hi,
Sure, I'd like to keep it low though...
Absolutely.
And this opposes our own design goal, we want max. Q for a given price.
I have a hard time believing that...Not only does it fly in the face of common engineering believe, it also contradicts everything else you correctly claim.
Doesn't make any sense to me. Not that I know anything about it ...
Designing for Class A operation is any designer's wet dream...it is sooo easy.
It relaxes most parameters except for costs.
Cheers,
Fred Dieckmann
Thanks for the academic explanation of excitation current but what are you saying about its influence on sound quality? Are you stating that it is not a significant parameter to compare? Are you saying that it can influence sound depending on its frequency/use profile? We are trying to determine if the Aq parameter will be useful in predicting sound performance so if you can, please make your point clearer.
Thanks for the academic explanation of excitation current but what are you saying about its influence on sound quality? Are you stating that it is not a significant parameter to compare? Are you saying that it can influence sound depending on its frequency/use profile? We are trying to determine if the Aq parameter will be useful in predicting sound performance so if you can, please make your point clearer.
Koinichiwa,
Yes. It goes hand in hand with running the core at lower flux. In effect you need to overdimension the transformers so that each can carry the full load, then using two in "low flux" configuration will halve the flux in the core.
Considering the fairly modest increase in price with rising wattage using a pair of 160VA transformers instead of a pair of 80VA transformers nowhere near doubles the cost.
And using instead one 320VA unit will NOT reduce the core flux, though it will be likely cheaper than the two 80VA units.
As usual - design is the art of making the right type of compromise.
If you wish to compromise sound quality line up with the buyers of $ 150 DVD/DVD-A/SACD/MP3 universal players made to the lowest cost in the far east.
Otherwise you have to accept the added cost. My (elsewhere) proposed 4 X 80VA Transformer supply for a "gainclone" using dual schottkies CAN be easily scald and applied to to the "low flux" connection.
The you have in fact a true dual mono supply with a current rating equivalent to 160VA, but core "headroom" and general handling equivalent to 320VA. Moreover, in Class AB Amplifiers the continous power handling of the maisn transformer is irelevant, as the quiescent current is fraction of "full scale" current.
More important is the behaviour of the transformer with varying loads - there I'd expect the low flux connection to really pull out the lap times and run rings around normal supplies. Of course, nothing that would match batteries, but obviously most people here are not adventerous, so be it.
Sayonara
Peter Daniel said:
Yes. It goes hand in hand with running the core at lower flux. In effect you need to overdimension the transformers so that each can carry the full load, then using two in "low flux" configuration will halve the flux in the core.
Considering the fairly modest increase in price with rising wattage using a pair of 160VA transformers instead of a pair of 80VA transformers nowhere near doubles the cost.
And using instead one 320VA unit will NOT reduce the core flux, though it will be likely cheaper than the two 80VA units.
As usual - design is the art of making the right type of compromise.
If you wish to compromise sound quality line up with the buyers of $ 150 DVD/DVD-A/SACD/MP3 universal players made to the lowest cost in the far east.
Otherwise you have to accept the added cost. My (elsewhere) proposed 4 X 80VA Transformer supply for a "gainclone" using dual schottkies CAN be easily scald and applied to to the "low flux" connection.
The you have in fact a true dual mono supply with a current rating equivalent to 160VA, but core "headroom" and general handling equivalent to 320VA. Moreover, in Class AB Amplifiers the continous power handling of the maisn transformer is irelevant, as the quiescent current is fraction of "full scale" current.
More important is the behaviour of the transformer with varying loads - there I'd expect the low flux connection to really pull out the lap times and run rings around normal supplies. Of course, nothing that would match batteries, but obviously most people here are not adventerous, so be it.
Sayonara
The reason I still didn't use batteries in my equipment is that I'm scheptical about the advantage they seemingly offer. It might be because I once read the review of Jeff Rowland preamp with option of either battery operating or not and the difference wasn't really great with good quality mains supply.
Another example of dissatisfaction with battery operation was Elso Kwak post at AA where he tried batteries on his DAC and didn't like them.
I was running my Sony DAT from battery as well and to tell the truth never really noticed substantial improvement in comparison to regulated supply.
My mains are not bad and living in susburbs has it's advantages. I have the utility transformer right beside my house and the connection goes directly to me. Not to mention 3 dedicated 30A, two fases lines, just for the stereo.
But I will try the battery with my gainclone and see if it's worthwile.
As to running split primaries in series at 120V, the reason I mentioned the power issues is because there was a lot of talk about it in Aleph X thread and many members claimed that it shouldn't change power rating of a transformer (and quite few were uncertain about it). Incidentally, I'm using that arrangement in my Aleph X, running 750VA transformer at half voltage.
Another example of dissatisfaction with battery operation was Elso Kwak post at AA where he tried batteries on his DAC and didn't like them.
I was running my Sony DAT from battery as well and to tell the truth never really noticed substantial improvement in comparison to regulated supply.
My mains are not bad and living in susburbs has it's advantages. I have the utility transformer right beside my house and the connection goes directly to me. Not to mention 3 dedicated 30A, two fases lines, just for the stereo.
But I will try the battery with my gainclone and see if it's worthwile.
As to running split primaries in series at 120V, the reason I mentioned the power issues is because there was a lot of talk about it in Aleph X thread and many members claimed that it shouldn't change power rating of a transformer (and quite few were uncertain about it). Incidentally, I'm using that arrangement in my Aleph X, running 750VA transformer at half voltage.
joensd said:
It's an electrostatic shield used to amoeliorate problems caused by interwinding capacitance (the capacitance between the primary and secondary windings). Interwinding capacitance allows high frequency garbage at the primary to couple to the secondary.
Interwinding capacitance charges/discharges into the local ground which can result in interchassis currents between components which ultimately manifests itself as noise at an amplifier's input.
se
Peter Daniel said:
Even a "perfect" mains supply still leaves you with ripple voltage on the rails unless you're using voltage regulation. And if you're using small value reservoir caps, your ripple voltage will be even greater. Further, the ripple voltage manifests itself as a sawtooth waveform so it will have significant harmonic content. And again, this is with a perfect, noise-free 60 Hz sinusoid mains supply.
I'm sure that Rowland preamp was using active voltage regulation so ripple isn't as much an issue as it would be in a non-regulated supply, which describes nearly all of the various Gainclones I've seen.
se
Tube output transformers for power
"I have a hard time believing that...Not only does it fly in the face of common engineering believe, it also contradicts everything else you correctly claim. Doesn't make any sense to me"
This transformer was being driven by a large power amp as a step up transformer. Coupling of noise from the primary was not an issue in this case. I am not sure what contradictions you mean. This was a large transformer and free from the presence of DC and very low frequency noise on the typical AC line. I really believe that that the wide bandwidth and low distortion core material might be beneficial for the higher frequency component of the diode currents charging the filter caps. Core material is a big deal in output transformers and if you think that the best grade steel is used for power transformers built for economy, you are more of an optimist than I am. I hope to do some measurements on the frequency spectra of the ripple current in the near future. I have measured the distortion of the AC line voltage and it not a pretty sight. There are a lot of things I and many others don't know about what makes a good rectified AC to DC power supply. That large differences in transformers, diodes, and filter cap choices make very substantial differences in how an amplifier sounds is a given at this point. I am very interested the reasons for these differences and methods for measuring and optimizing these factors. I think there has been little research into this area that most engineers find boring and unchallenging. I, on the other hand, still find that there are many questions on why power supplies sound so different and what design choices and measurements are most important for building good sounding supplies.
'I'm sure that Rowland preamp was using active voltage regulation so ripple isn't as much an issue as it would be in a non-regulated supply, which describes nearly all of the various Gainclones I've seen."
Mine has a LT1084 regulated supply.........
"I have a hard time believing that...Not only does it fly in the face of common engineering believe, it also contradicts everything else you correctly claim. Doesn't make any sense to me"
This transformer was being driven by a large power amp as a step up transformer. Coupling of noise from the primary was not an issue in this case. I am not sure what contradictions you mean. This was a large transformer and free from the presence of DC and very low frequency noise on the typical AC line. I really believe that that the wide bandwidth and low distortion core material might be beneficial for the higher frequency component of the diode currents charging the filter caps. Core material is a big deal in output transformers and if you think that the best grade steel is used for power transformers built for economy, you are more of an optimist than I am. I hope to do some measurements on the frequency spectra of the ripple current in the near future. I have measured the distortion of the AC line voltage and it not a pretty sight. There are a lot of things I and many others don't know about what makes a good rectified AC to DC power supply. That large differences in transformers, diodes, and filter cap choices make very substantial differences in how an amplifier sounds is a given at this point. I am very interested the reasons for these differences and methods for measuring and optimizing these factors. I think there has been little research into this area that most engineers find boring and unchallenging. I, on the other hand, still find that there are many questions on why power supplies sound so different and what design choices and measurements are most important for building good sounding supplies.
'I'm sure that Rowland preamp was using active voltage regulation so ripple isn't as much an issue as it would be in a non-regulated supply, which describes nearly all of the various Gainclones I've seen."
Mine has a LT1084 regulated supply.........
XFORMER-FILES.
Hi,
That is quite a different thing altogether.The xformer provides for a galvanic isolation of the amp/supply preceding it.
As such this would be beneficial indeed.
Here you seem to advocate the use of high qualility core material, as for an OPT, in powerxformers.
IMHO this would not be beneficial and that is what I meant when I mentioned the contradictary demands for the two types of tasks.
If we were to do that we would face the same problems we have with toroids: too much permeability lets too much noise through.
We should also consider the use of chokes in the PSU IMO.
it is my believe that these can be put to very good use in PSU design, especially for improvements in noise rejection rates.
Cheers,
Hi,
That is quite a different thing altogether.The xformer provides for a galvanic isolation of the amp/supply preceding it.
As such this would be beneficial indeed.
Here you seem to advocate the use of high qualility core material, as for an OPT, in powerxformers.
IMHO this would not be beneficial and that is what I meant when I mentioned the contradictary demands for the two types of tasks.
If we were to do that we would face the same problems we have with toroids: too much permeability lets too much noise through.
We should also consider the use of chokes in the PSU IMO.
it is my believe that these can be put to very good use in PSU design, especially for improvements in noise rejection rates.
Cheers,
Okay guys, let's keep our focus on transformers alone. We can start discussing inductance coils and other passive part strategies after we pinpoint and assign a weight of importance to the issues in the transformer, agreed? Now let's review.
The issues introduced so far and their order of importance are:
1. core flux density (lower is better)
2. DC blocking/inhibiting (less DC is better)
3. quiescent current (lower Aq is better)
3a. quiescent current (lower under load is better and standing Aq measures are meaningless)
4. conductor thickness (thicker sounds better)
Are we all in accordance with the order of importance of this list? Please add any other issues or edit the order to reflect your opinions. Once our list is complete, we will discuss each issue in greater detail and begin assigning a weight of importance to each.
The issues introduced so far and their order of importance are:
1. core flux density (lower is better)
2. DC blocking/inhibiting (less DC is better)
3. quiescent current (lower Aq is better)
3a. quiescent current (lower under load is better and standing Aq measures are meaningless)
4. conductor thickness (thicker sounds better)
Are we all in accordance with the order of importance of this list? Please add any other issues or edit the order to reflect your opinions. Once our list is complete, we will discuss each issue in greater detail and begin assigning a weight of importance to each.