Post by foveaux on Sept 13, 2019 17:02:04 GMT 12
Some while ago I read, on the net, that current mode output CD players, headphone amps and amplifiers had 'advantages' for music replay cf the usual voltage mode output. I can't find that article again, but I'm interested to learn about this, at the overview level. Would appreciate some feedback from colleagues on their experiences and/or point me to further reading, recommended current mode output equipment, etc. Also happy to be told 'stop barking up the wrong tree'  Thanks in advance chaps.
"I see music as a lifetime affair." [Rory Gallagher]
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Post by Owen Y on Sept 13, 2019 17:27:12 GMT 12
Nelson Pass wrote about the subject of Current Source Amplfiers on his First Watt site & why such amp designs particularly suit the category of 'Full-range' loudspeaker drivers. Possibly is this the article that you refer to?
|
Post by foveaux on Sept 13, 2019 17:49:50 GMT 12
Thanks Owen, must have been that, as I followed the audiophilliac (guttenburg) interviews and dived off on to tangents from that. I recalled a number of current source headphone amps too. So little current mode output equip, is it too esoteric, unproven? Am I right - that indicates it has only narrow specialised application?
"I see music as a lifetime affair." [Rory Gallagher]
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Post by colinf on Sept 13, 2019 20:15:25 GMT 12
Current mode output amps are probably most suited to speakers than preamps and CD players etc. They have high output impedance and low damping factor which changes the total damping on speaker cones to be less damped. Speakers need an ideal damping effect on the cone or the frequency response can be affected. Ideally a Qts of 0.7...a bit higher or a bit lower as the speaker designer wants the speaker to work best with the room. Some highly efficient speakers have an efficient magnet system and can become overdamped by using them with a voltage mode amp, with its low output impedance. That can make the bass roll off prematurely. Most if not all solid state amps have very low output impedance and tend to be unsuitable for highly damped cones. OTOH inefficient speakers tend to need all the damping factor (and thus low impedance) from a voltage source amp they can get. So it depends on the speaker really. But in the middle and about right for damping factor for efficient-ish speakers are single ended amps with no negative feedback (NFB). No NFB amps tend to have a damping factor of around 1or 2. Most NFB amps have a damping factor over 25 and can be as high as 1000! For CD players, preamps and source equipment it’s usually better to have a voltage source output so that it can drive the capacitance of interconnects successfully and partly absorb RF. But even source equipment ideally should have a defined output impedance to work best with the specific interconnect you’re using, otherwise you get reflections and resonances and/or rolled off response at RF frequencies. RF interacts with the non-linear bit of the source equipment’s output impedance and can add distortion. In a circuit inside a box though, the circuit is shielded from the potential for RF problems from outside the box and current mode can be more linear, simply from the fact that there are fewer stages needed to amplify a signal. Then there is current mode input for MM phono stages. The input impedance is made purposefully low which rolls off the response from the cartridge and lessens the effect of the cable capacitance from the turntable. and is compensated for in the RIAA equalisation in the circuit. That system has the effect of extending the high end frequency response with less phase shift than a normal 47k input impedance.
AMR-iFi R&D
|
Pundit
Post by garym on Sept 14, 2019 11:32:28 GMT 12
What Colin said. My 2c. Most loudspeakers are designed on the assumption they will be fed from a voltage source. The behaviour of the crossover network needs to make this assumption as does the acoustic design of enclosure / driver. As regards amplifier damping factor, this is dominated by the resistance in both the speaker cable and any crossover coil in series with the woofer. There could easily be half an Ohm there and that limits the damping factor with an 8 Ohm driver to 16. Some loudspeaker manufacturers design specifically for non flat behaviour. For example I've seen Zu state that their speakers need high output impedance amplifiers to get the (undamped) bass response they intend. This makes amplifiers selection a lottery. Current output is uncommon because of all of this.
|
Post by foveaux on Sept 14, 2019 11:48:16 GMT 12
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Post by foveaux on Sept 14, 2019 12:07:24 GMT 12
What Colin said. My 2c. Most loudspeakers are designed on the assumption they will be fed from a voltage source. The behaviour of the crossover network needs to make this assumption as does the acoustic design of enclosure / driver. As regards amplifier damping factor, this is dominated by the resistance in both the speaker cable and any crossover coil in series with the woofer. There could easily be half an Ohm there and that limits the damping factor with an 8 Ohm driver to 16. Some loudspeaker manufacturers design specifically for non flat behaviour. For example I've seen Zu state that their speakers need high output impedance amplifiers to get the (undamped) bass response they intend. This makes amplifiers selection a lottery. Current output is uncommon because of all of this. Thanks Gary, Very grateful for your reply. Noted that current output is uncommon, and therefore a 'step too far' for me, an average audio punter. I really need to stop reading audio threads late at night, where I convince myself I've discovered the audio holy grail!
"I see music as a lifetime affair." [Rory Gallagher]
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Pundit
Post by belbo on Sept 14, 2019 15:44:43 GMT 12
Do you remember the name of the Japanese brand? A few produce small full range speakers (usually up to 4 inches), including Airtight, Audio Tekne, 47 Labs, FAL makes a range of very unique speakers that work full range.
|
Post by colinf on Sept 14, 2019 20:24:17 GMT 12
Basically, if your speaker has a crossover in it, it’s probably best to drive it from a voltage-source (ie conventional) amp. The design of the crossover is intended, as GaryM mentioned, for that kind of amp. If it’s a full range speaker without a crossover it’s game on for whatever amp type works best! I think the Questyles have a headphone output which is current-mode..? That would work fine for headphones as most have a single driver. Most headphones (but not all) seem to be designed for a source impedance of 150 ohms, especially 600 ohm Sennheisers etc. Hope this helps!
AMR-iFi R&D
|
Post by foveaux on Sept 14, 2019 22:27:35 GMT 12
Thanks again Colin Very helpful. Yes, current mode amp is clearly a non-starter for my second system ELAC speakers. I was presuming 'current mode' was a newly emerging technology, like class D etc. has been in recent years. Great to have your, Gary and Owens expertise avail on our forum.
"I see music as a lifetime affair." [Rory Gallagher]
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Pundit
Post by garym on Sept 15, 2019 12:32:22 GMT 12
Colin... Not sure I agree about full range drivers and current mode. The acoustic design for low frequency performance is still dependant on theil small parameters, box volume etc and source impedance... Ie low impedance voltage source amps... Unless maybe you design a one off combination of current mode / driver / box design. As can be seen in Stereophile measurements, an amp with non zero and variable output impedance will interact with a loudspeaker to give unpredictable results. Eg a Set amp will generally behave as a tone control...
|
Pundit
Post by garym on Sept 15, 2019 12:34:02 GMT 12
Footnote. Krell used to have "cast" connection between pre and power etc... This was a current mode connection specific to their products.
|
Post by foveaux on Sept 15, 2019 13:01:01 GMT 12
"Free - I miss that band, but when I look back, we were very young" [Paul Rodgers]
848 posts
|
Pundit
Post by belbo on Sept 15, 2019 18:34:53 GMT 12
|
Post by sub on Sept 15, 2019 19:26:18 GMT 12
Not being technically inclined, I have limited knowledge of what is being discussed here! However, searching on line for a review of a Sony TA-F444ES amp, I found a review that said, it like many ES designs, used a technology in Sony speak called “audio current transfer” (ACT) where current and not voltage is transferred from one stage to another. Is this what is being discussed here? www.highfidelityreview.com/inside-out-sony-ta-f444es.html
|
Post by colinf on Sept 15, 2019 19:30:26 GMT 12
Colin... Not sure I agree about full range drivers and current mode. The acoustic design for low frequency performance is still dependant on theil small parameters, box volume etc and source impedance... Ie low impedance voltage source amps... Unless maybe you design a one off combination of current mode / driver / box design. As can be seen in Stereophile measurements, an amp with non zero and variable output impedance will interact with a loudspeaker to give unpredictable results. Eg a Set amp will generally behave as a tone control... I agree Gary, ported and vented boxes are designed for voltage amps as you don’t want to pump the current from a current mode amp through the resonant peaks in the speaker’s impedance graph at bass frequencies. Varying the impedance drive from the amp would act as a tone control, as you say. Some people like that though, within limits. If it’s an open baffle or sealed box, ie, one with a nice smooth impedance graph, you can get away with an amp with higher output impedance, approaching current drive.
AMR-iFi R&D
|
Pundit
Post by garym on Sept 15, 2019 20:14:42 GMT 12
Sub... Sounds like Krells CAST mode...
|
Post by sub on Sept 18, 2019 10:03:29 GMT 12
I enjoy reading this technical stuff, even though it can be hard to understand.
Here, I am somewhat puzzled by the notion of current without voltage. Had to do a bit of basic learning about electricity when we bought our off grid home (solar with battery storage) 9 years ago.
I learnt that current is the flow of electrons in a conduit, usually a copper wire, that the size of the current is measured in amperes, that the force pushing the current is measured in volts, and that the amount of power used by a device connected to the wire is measured in watts (or joules). From that it seems that current cannot flow unless there is a force (volts) pushing it?
So, how can current pass without voltage? How can they be separated, as is being discussed here, where it is implied that an amp can deliver either current or voltage?
Are the laws of electricity different when it comes to audio devices and similar? For instance, the output power of an audio amplifier is usually quoted as so many watts per channel, Whereas watts is usually a measure of the amount of power consumed, rather than delivered!?
|
Pundit
Post by garym on Sept 18, 2019 10:30:47 GMT 12
Current is flow of electrons in coulombs per second (amps). Voltage is energy per unit charge in joules per coulomb (volts) . Power is energy consumed per unit time or joules per second (watts). You can rearrange these in simple math equations to good use. Also Ohms law... Volts = current times resistance. Current = Volts divided by resistance. Etc You can see from this that higher current flows when more voltage is available or when resistance is reduced. If you use current to carry the signal you are usually talking about low resistance circuits where the voltage will be low (but not zero). Likewise in the more usual voltage transfer of signal you are talking about high resistance and low (but not zero) current. Whew...
|
Post by Owen Y on Sept 18, 2019 10:53:53 GMT 12
sub - I like to think always in terms of Ohm's Law:  So - the greater the Resistance, the greater the Voltage developed or required & greater the Amps. The lower the Resistance, the lower the Voltage & higher the Amps. And so on... Most amplifiers are designed to be connected to lspkrs which have a certain amount of Resistance (say 4/8/16 Ohms). And (need to) develop a certain amount of Voltage in order to 'drive' such a lspkr. Some lspkrs have very low Resistance (either by design or type or over some part of their freq range, eg a single driver run 'full range' will show a 'resonant freq' down low where its Resistance plummets) - amplifiers to drive such lspkrs need to be able to deliver mucho current because of this dropping Resistance load down at LFs (Ohm's Law). So, there is always this relationship between Voltage-Restance-Current. The terms 'voltage amplifier' or 'current amplifer' are really just semantics or terminology. (Although different amp cct topolgies/devices might be better at delivering voltage vs current, or vice versa.) (I hope I haven't got anything too badly wrong here, colinf, garym ) This might help 
|
Post by sub on Sept 18, 2019 12:04:37 GMT 12
Whew! indeed. Think I’ve got it now, thanks Gary and Owen. So, in such devices the current flows with minimal voltage, not zero volts.
Ohms law is something I’ve spent time trying to understand, as the tech guys, on the solar power forums I hang out on from time to time, bring that up fairly regularly. That cartoon I’ve seen before, but it does help in getting your head around it!
|
Post by colinf on Sept 18, 2019 19:03:25 GMT 12
I usually just like to visualise it by thinking of a hose with water flowing through it. The pressure is equivalent to voltage. The flow is equivalent to current. The diameter of the hose is equivalent to resistance. 1 amp flows through a 1 ohm resistor at 1 volt. The resonance of a speaker at its resonance frequency results in an increase in impedance. That’s because the cone’s velocity is higher at that frequency and acts more like a generator (called back EMF). That works in opposition to the incoming current flow and less current will flow at that frequency, for the same voltage input as other frequencies. A current drive amp will attempt to drive the same current at the resonance frequency as at other frequencies and there is a resulting peak in acoustic output at the resonance frequency. So current drive amps can act to make bass lumpy or uncontrolled, especially for woofers in ported boxes. Where a voltage amp delivers more current to a speaker is when the voice coil is being driven past the end of the magnetic gap and the resistance reduces dynamically, demanding lots of current from the amp. A current drive amp in the same situation will result in higher distortion. That can happen with speakers with a short voice coil gap, or with subwoofers reproducing the sound of thunder.
AMR-iFi R&D
|
Member
Post by Wonfor on Mar 1, 2021 19:03:07 GMT 12
Thanks Colin,
Most interesting, I use current FBK on the pre-out and voltage to keep the ref zero at zero.
On SECA designs I use on the pre drive the Class A driver single ended and close current FBK has the poor little transistor will be driving a large MOSFET and overall Voltage FBK on the DC output, all done single ended and DC coupled.
MC stage I use my patented constant current sliding bias design which use Kirchoff Law.
MM I use VFBK low noise component and devices that can run at -40C on the top designs, I then cool them down using a heat pump or a Peltier Heat Exchanger. This reduce the noise floor very well. But a pain to do.
On my large amps I use twin FBK and remote sense to the speaker, I treat them as large industrial design PSU's with a dynamic output and using the audio in as the voltage reference. I also design them to have very low output impedance in the order of mR and the DC offset is lower than 100uV. But the dT on rise and full time is about 10nS positive and 12nS negative at a max current of 100Amps, So the internal PSU i.e. the DC +V & -V have large storage and fast delivery.
All great fun.
Today I start a new idea using some SiC devices for normally RF work. The price of these toys is crazy.
Best Col
|
Post by cartridgeguyonline on Mar 1, 2021 20:53:30 GMT 12
... and straight through to the keeper.
|
Post by colinf on Mar 1, 2021 21:46:52 GMT 12
Nice, the sliding bias arrangement for MC sounds interesting. Kirchhoff’s law...all current going into a node is the same going out, and the total charge around a circuit loop will be net zero, true for circuit sizes less than the wavelength it’s handling, How are you applying this to MC? I once made a class G circuit, single ended with the power supply driven by the audio signal. The output impedance was then more linear so the input stage’s nfb function was more constant. MM with a peltier effect module to cool it... but the noise is limited by the cartridge source impedance, and s/n ratio with records is limited anyway?
AMR-iFi R&D
|
Member
Post by Wonfor on Mar 1, 2021 22:12:50 GMT 12
seca.freeforums.net/The above link will take you to stored files for anybody to look at, on there somewhere is the MC circuit I am writing about plus a few more boring things. Circuits for Claymore's are there also for those who need it. Catch up later Maria wants to use my body, I am 6ft4in ideal for cleaning shelve she can't reach as she is 5ft, a ideal size to fit in suit case for a cheaper holiday. (do not tell her) he he
|
Pundit
Post by peter0c on Mar 2, 2021 10:16:48 GMT 12
Whew! indeed. Think I’ve got it now, thanks Gary and Owen. So, in such devices the current flows with minimal voltage, not zero volts. Ohms law is something I’ve spent time trying to understand, as the tech guys, on the solar power forums I hang out on from time to time, bring that up fairly regularly. That cartoon I’ve seen before, but it does help in getting your head around it! Current amps have a constant output voltage - it is the output current that varies with the source input voltage - for example my First Watt F1J has a constant 13.6 volts output. The same as a fully charged car battery which is an amusing coincidence. !
|
Post by colinf on Mar 2, 2021 20:50:28 GMT 12
But why did Maria allow the use of shelves that were too high, knowing she couldn’t get to them to clean them? Just looked at the MC circuit, emitter input, how it should be. So no phase inversion and low noise at 1mA. I’d possibly add a servo circuit for one of the power supplies to keep the input DC at 0v, so no current flows through the MC coils. SiC...for output stages? I think Nelson Pass had a go with Semi South devices I fiddled a bit with Sony Vfets a while back.
AMR-iFi R&D
|
By Mr Kazutoshi Tsukahara (had link with 47 Labs)
