This is a very good point. What happens with an ordinary amplifier-cable-speaker combination will depend on the characteristics of all 3 when we are talking about stability and/or EMI ingress. There could be a "perfect storm" of conditions. It is also certainly true that some amplifiers, be they ordinary fidelity or high-end audiophile, are not designed with adequate stability for all load situations. Their designers may be pushing the envelope on feedback compensation, knowingly or un-knowingly.
There are also amplifiers out there where designers have chosen to omit the outpul series R-L network for fear of the series inductor degrading the sound quality. These amplifiers may be more susceptible to instability problems. It may even be the case that amplifiers that are diligently tested with a variety of direct load capacitances from, say, 1000 pF to 1 uF may still have some stability issues under more complex real-world loads. Even the use of a 1 uH + 2.2 ohms series R-L network can sometimes make a world of difference in stability under unusual conditions.
We all know that emitter followers hate direct capacitive loads, and some CFP output stage designs might even be more susceptible to instability under capacitive loading.
Modern output transistors are also much faster than in the past, with ft in the 20-60 MHz neighborhood. These fast transistors significantly help to reduce high-frequency and switching distortion in the output stage, but emitter followers built with these probably hate direct capacitive loads even more.
Even if phase margin is adequate, it is crucial to pay just as much attention to gain margin. For example, if one uses a lead capacitor in the feedback network to improve phase margin, if not carefully implemented this can degrade gain margin.
One strenuous test for stability is to cut the closed-loop gain in half and then evaluate stability. If oscillation or instability on a portion of the waveform does not occur for any signal level into a resistive load, you know you've got about 6 dB of gain margin (we often want more). This test can be made more strenuous by doing the above test with the usual variety of direct capacitive loads.
Cheers,
Bob
Isolation Transformer
Slightly off-topic. Has anyone used a toroid power transformer with the usual 2-winding primary as a line isolation transformer by connecting one primary to the line and the other primary to the load?
In such a case, I'm guessing that its capability would be on the order of half its normal rated VA, since only half of the usual primary is being used for line input, assuming that copper losses are the governing constraint.
In a similar context, might a conventional non-toroid transformer with dual primary windings provide a bit more isolation? Most times I've measured primary-secondary capacitance, the toroid of similar VA has had more capacitance, but I have never measured capacitance between the two primaries of either type of transformer.
Thoughts?
Cheers,
Bob
Slightly off-topic. Has anyone used a toroid power transformer with the usual 2-winding primary as a line isolation transformer by connecting one primary to the line and the other primary to the load?
In such a case, I'm guessing that its capability would be on the order of half its normal rated VA, since only half of the usual primary is being used for line input, assuming that copper losses are the governing constraint.
In a similar context, might a conventional non-toroid transformer with dual primary windings provide a bit more isolation? Most times I've measured primary-secondary capacitance, the toroid of similar VA has had more capacitance, but I have never measured capacitance between the two primaries of either type of transformer.
Thoughts?
Cheers,
Bob
It contains useful basic informations (hints) to the real world problems when using amplifier/cable/loudspeaker combinations and some about the possible bias effects as well.
But it contains one of the basic logical fallacies as well; misleading listeners in a so-called listening test is usually just an easy task but it does not allow conclusions about listening abilities under different conditions.
If in doubt about that, do the same with loudspeakers of different sizes/appearance and just pretend to switch while not changing the loudspeaker in reality. The results will be the same - as you still haven't changed anything - but which conclusion will be justified?
The interesting point about Dunlavy's cable post was to me, that he later (iirc) in other posts/papers was adamant that a perfect impulse response in a speaker will lead to audibly better music reproduction, but never (again afair) offered any corrobation by controlled listening experiments.
Have you seen the article by Newell/Holland where they mentioned some surprising effects seen by Voishvillo and others when combining an amplifier to the same loads via cables of different contructions?
https://www.eetimes.com/document.asp?doc_id=1274851
No frequency response measurement but multione FFTs.
For my ground loop article, I measured the primary to secondary capacitance of a 1.2kVA transformer at 1.3 nF and on the 500 VA at 500 pF.
When the transformer interwinding screen was grounded, the interwinding capacitance went to <100 pF.
I also measured the common mode and series mode BW of both transformers with 50 Ohm source and 50 Ohm load. In series mode (normal transformer mode in other words) the BW was 60 kHz on both of them, while the CM BW was at the limit of my generator which is 3MHz.
Both transformers were Toroids.
Hello All,
While I am still vertical I will run some of these plots on my bench. If the pacemaker does not zap me I will post the results. Laugh!
Next weekend there will be time in the lab.
I am also interested in the effect on impedance and phase plots with different amplifiers and conductors.
I have a couple of different amplifiers to test; Parasound and Crown. Plus wire types; SO cord, "speaker" zip cord and 12AWG THHN off the roll.
Thanks DT
On the bench is a APx555 and APx1701.
While I am still vertical I will run some of these plots on my bench. If the pacemaker does not zap me I will post the results. Laugh!
Next weekend there will be time in the lab.
I am also interested in the effect on impedance and phase plots with different amplifiers and conductors.
I have a couple of different amplifiers to test; Parasound and Crown. Plus wire types; SO cord, "speaker" zip cord and 12AWG THHN off the roll.
Thanks DT
On the bench is a APx555 and APx1701.
I have a transformer like this on my bench but just use it to power a scope for making current measurements on a .01 Ohm resistor in the AC line.
About 6 nF between primaries! Primary secondary 200 pF.
Bob, yes I have thought about it but not tried it.
Would this arrangement not act as a current transformer with the unused secondary generating high voltages if not loaded ?.
Along this line my thoughts also included connecting R or RC (RCL?) networks loading the unused winding (sorta like a telephone hybrid), anybody have thoughts on this ?.
I have also thought about such loading/damping networks across the output winding of a current transformer monitoring the AC supply Active wire, anybody have thoughts on this also ?.
Dan.
A current transformer has only very low voltages across any winding as the secondary is shorted. You must never connect a current transformer across a voltage source. Current transformers need far fewer turns as they don't need so much inductance to deal with small voltages.
If you use the primaries at mains voltage, the secondaries will be at their design voltage. Its a voltage transformer because you haven't shorted out any winding and you have connected it to a voltage source.
If you use the primaries at mains voltage, the secondaries will be at their design voltage. Its a voltage transformer because you haven't shorted out any winding and you have connected it to a voltage source.
That seems safely conservative. Actually, since you have near zero heat in the secondaries, you can go more than half. But not full VA since the cool secondary does not fully compensate the hot primary. And the in-between depends on details. 66% is probably reasonable for most cases. You also ask: what is the load? Speech/music B/D amps idle low power and the peaks do not count for much. But with smaller iron you may have enough sag on peaks to "matter". Steady "full" (66%) load like heaters, class A, or B/D testing, keep a finger on the transfo and be aware if it gets hot. (Most good torroids will take more heat than a finger, at least for the length of a listening or test session.)
But what is "isolation"? The two 120V windings are expected to be connected together one way or the other. So maximum 240V end to end, and 120V through insulation if they aimed for minimum stress. So yeah to "float" a grounded wall-outlet to a grounded test bench you can reduce to the interwinding capacitances instead of galvanic connection. But sometimes you find folks trying to float to 1,000+V potentials. While a good PT has been hi-potted for thousands of volts for a minute it may not be long-term safe up there.
No, this is not a current transformer. The secondaries will idle at normal NO-load voltage.
The primary winding pairs would be bifilar wound and equal in length so one could imagine these in the sense of parallel capacitor plates. I doubt this use would comply with safety regulations.
If as suggested there can be as much as 6nF in this sort of configuration emi is still going to get through.
If as suggested there can be as much as 6nF in this sort of configuration emi is still going to get through.
These are good points. I was also wondering if the two primary windings were bifilar would. That certainly makes sense from an intuitive engineering point. It would also explain the large capacitance seen between the two primary windings. Is it expected that all constructions of such transformers would have bifilar-wound dual primaries (e.g., both toroid and EI constructions)?
A pragmatist might argue that bifilar winding of the two primaries is not strictly required, since the number of turns can be made identical and that maybe the slight resulting differences in resistance might not matter much. The difference in resistance would likely cause one primary to carry a bit more of the load than the other when they are parallel-connected, but that might not end up being significant enough to justify the extra cost of implementing bifilar windings.
I guess that for any given transformer we could get a clue by comparing the resistances of the two primaries and see whether they are nearly identical or slightly different. I am clearly not a transformer expert
.Cheers,
Bob
These are good points. I was also wondering if the two primary windings were bifilar would. That certainly makes sense from an intuitive engineering point. It would also explain the large capacitance seen between the two primary windings. Is it expected that all constructions of such transformers would have bifilar-wound dual primaries (e.g., both toroid and EI constructions)?
A pragmatist might argue that bifilar winding of the two primaries is not strictly required, since the number of turns can be made identical and that maybe the slight resulting differences in resistance might not matter much. The difference in resistance would likely cause one primary to carry a bit more of the load than the other when they are parallel-connected, but that might not end up being significant enough to justify the extra cost of implementing bifilar windings.
I guess that for any given transformer we could get a clue by comparing the resistances of the two primaries and see whether they are nearly identical or slightly different. I am clearly not a transformer expert
Cheers,
Bob
Neither am I.
Someone might know if it is necessary to match the core material and mass so an isolation transformer works in sych with the one to be isolated.
There is some remenence in steel when it is magnetised so I am unclear how different transformers of different mass etc would behave in the time domain and with back emf from the isolated transformer.
That might not matter for a pre-amp supply however it could be disaster if the supply capacitor charging currents have missed the bus at the top and bottom ends of the 50/60Hz charging cycle.
No.
Not necessary (the DCR is tiny compared to working impedance, and a 0.98:1.02 division does no harm). Bifilar is more prone to varnish breakdown; why risk a flaming lump for no real benefit?
The mechanics of winding a toroid *may* encourage bifilar to reduce the money-motion of winding. In EI the two windings will almost always be one over the other.
Hello All,
At first look you may think that it is volt amps or power that is the limiting factor. A little closer look will show that it is the wire gauge that limits current and increasing current through a fixed size conductor that causes heat. In addition to increased current caused heat you get flux saturation with flat top output voltage curves, with a spray of high frequencies. The opposite effect you were looking for.
Hospitals are full of power conditioners most have isolation transformers similar to TriadMagnetics N-68X. The best isolation / power conditioners are motor driven generators.
Thanks DT
Was it Nordost?
I have also often been skeptical as well. Probably a little more skeptical of power cords than speaker cables. If the role of the power cord is to suppress RFI/EMI ingress, I would think that a good component would already have adequate line filtering, but maybe they don't. Apart from reduction of RFI/EMI ingress, it is hard for me to understand how the last 6 feet of line voltage delivery to the unit could make much difference, when it is usually preceded by up to 50 feet of 12 or 14 AWG Romex. Maybe we need a Zobel network on the line
Cheers,
Bob
Nordost sounds familiar, but there are so many names and rooms to remember in a short time at such an event.
I would very much welcome a double blind test if they were able to arrange it. The difficult part is the time lapse. They have to stop the music, power down the CD player, swap the cord, and restart, so it can't be a sudden switchover. Also, being a skeptic, I was expecting that there would be no difference in sound.
The late Peter Aczel (The Audio Critic) covered the issue of "high quality" line cords after miles of "substandard", non audiophile wiring and he, also, dismissed fancy cables as useless.
My alternate solution to line cords being priced similar to college tuition would be to not have a line cord. Much less expensive to run the CD player from a battery. At the end of each CD, swap the battery with the other one in the charger. Pure DC. Now, what about your "special" line cord?
As for interconnecting cables to upgrade to something expensive, how many interconnects are there? Just a few between boxes. Within the box we have hundreds of non audiophile solder joints, bonding wires, foil paths, mechanical pot, switches, relays, etc. What one does need to watch for these days is cheaper cables that are a twisted pair rather than shielded. They are depending on low impedance of modern devices to reduce the hum.
Last edited: