? You have a good point. Something is wrong with the Edcor specs. Edcor has 2 different 1:5 mic input transformers, one having a spec of 0.3H Lpri and the other having a spec of 1.2H Lpri. I don't see how either of these delivers the frequency response shown in the pretty AP plot at 150 ohms or greater source impedance... They also give "Primary impedance" on the order of 2.6K for one and 12K ohms for the other 😕
Getting back to the original question. I have found that physically smaller transformers generate more low frequency distortion (measured) than large ones. Driving impedance, secondary load and amplitude all being equal. So don't use a mic transfo for line level.
But I wonder about sreten's point of a higher load impedance meaning less current and perhaps less distortion and extended low end response. To me that seems most germane to the OP. Can one cheat a bit with a lesser load?
The answer may be somewhere in the proceeding 4 pages, but I don't see it. Thanx.
But I wonder about sreten's point of a higher load impedance meaning less current and perhaps less distortion and extended low end response. To me that seems most germane to the OP. Can one cheat a bit with a lesser load?
The answer may be somewhere in the proceeding 4 pages, but I don't see it. Thanx.
Cheaters never prosper.
The cheat trades off ringing for current. If you Zobel to calm down the ringing, the current at treble frequencies increases since the Zobel reduces the load there.
The cheat trades off ringing for current. If you Zobel to calm down the ringing, the current at treble frequencies increases since the Zobel reduces the load there.
Regardless of higher impedances, the CMMI-5C is rated for a full bandwidth for signal levels up to 10dBu. My source is a CDP, so 2V out, ~8.2dBu, so I'm not in the red yet.
The spec sheet graph show that for a 20Hz input of 10dBu, distortion is ~0.8%. I think that's pretty good considering most music's low ELF content, which would not likely be at full scale output anyway.
IG
Last edited:
We talked about bandwidth issues, distortion, but one thing that has not been brought up and that I could see being of concern with higher impedances would be noise. What do you guys think?
IG
IG
The answer is no. There is no substantial tradeoff allowing one to cheat Bmax by reducing the load current. Please see Faraday's law again. Smaller cores have less current carrying capacity, just like wire, but the voltage/frequency saturation occurs independent of any load at all.
SY has a point about the ringing due to undertermination, but I'm trying to address the notion that voltage headroom can be gained by reducing the load current... I don't think so, and my attempts as a youngster to cheat this on power transformers were rewarded with a release of smoke. When I learned Faraday's law, I understood why.
If I'm somehow all wet, I wish someone would educate me.
Cheers,
Michael
PS I think the original question is also answered, but the S/N ratio is low ;-). THe CDP probably has a low enough Zout to drive the inductance to some reasonable LF limit; measurement would confirm this. Likewise one might expect some HF peaking due to undertermination but maybe a Zobel would tame it. THe signal voltage seems to be within the transformer spec. so while not the ideal transformer, it could be satisfactory. I wouldn't invest in these transformers for the purpose stated, but if they are on hand and sound OK, why not use them? Of course, sometimes after measurement the sound degrades...
Last edited:
I did buy them a bit hot-headedly for another project for which I ended-up finding them unsuitable. I could have worked them in, but it would have gone against the whole premise...
For the technically inclined, sound quality is both good and bad before taking a measurement....
Cheers,
IG
Last edited:
Hi,
I find this somewhat confusing from the numbers in the original post.
OK the load is ~ 39k, that for a primary load of ~1.5k, presumably
to be driven at a 10:1 ratio by a source impedance of - 150 ohms.
(the load being a 10:1 ratio of the 150:3.9K ratio of the trafo).
The amplifier impedance is ~390k, implying a primary load of ~ 15k.
implying it can be driven by source impedance of around ~ 1.5k.
Somehow one suspects it will not really work this way 10 times spec.
The simplest analogy one could make is I would suggest using
the equivalent of ten times thinner wire on both sides for ten
times the windings on both sides of the same core and all would
be hunky dory ? That in fact would be the square equalivent, i.e.
using x 3.162 as the windings change, or I'm getting confused ?
But wouldn't this only be true for the same nominal voltage levels ?
It cannot be true AFIACT if you want higher nominal voltage levels.
How do you change the design for the same core for higher voltage
levels ? AFAICT you reduced the numbers of the windings ......
and for the same power throughput this implies higher R loadings.
As its a CDP and passive pre-amp and assuming a 10k pot the
maximum source impedance is around 2k as I recall, so there
mightl be some interaction with the primary as can be < 10:1.
It really cannot be that difficult can it to describe what is going on
in simple terms and how well it will work for this particular case ?
As it seems the transformer can handle far higher voltage levels
than you get from a microphone (surely there is some trade off
here* ? Trafos for MC phono cartridges IME are tiny) then is it
not as a first question what is the likely best secondary load
given the source impedance is 2K maximum ?
Is there anything that means it simply will not work well ?
Are there any real bass issues ?
Are there any real treble issues?
rgds, /sreten.
* For MC cartridges I understand you need to make the transformer as
small as possible for the given maximum levels such that very small signals
are not swamped by the high magnetisation issues of an overlarge core.
I find this somewhat confusing from the numbers in the original post.
OK the load is ~ 39k, that for a primary load of ~1.5k, presumably
to be driven at a 10:1 ratio by a source impedance of - 150 ohms.
(the load being a 10:1 ratio of the 150:3.9K ratio of the trafo).
The amplifier impedance is ~390k, implying a primary load of ~ 15k.
implying it can be driven by source impedance of around ~ 1.5k.
Somehow one suspects it will not really work this way 10 times spec.
The simplest analogy one could make is I would suggest using
the equivalent of ten times thinner wire on both sides for ten
times the windings on both sides of the same core and all would
be hunky dory ? That in fact would be the square equalivent, i.e.
using x 3.162 as the windings change, or I'm getting confused ?
But wouldn't this only be true for the same nominal voltage levels ?
It cannot be true AFIACT if you want higher nominal voltage levels.
How do you change the design for the same core for higher voltage
levels ? AFAICT you reduced the numbers of the windings ......
and for the same power throughput this implies higher R loadings.
As its a CDP and passive pre-amp and assuming a 10k pot the
maximum source impedance is around 2k as I recall, so there
mightl be some interaction with the primary as can be < 10:1.
It really cannot be that difficult can it to describe what is going on
in simple terms and how well it will work for this particular case ?
As it seems the transformer can handle far higher voltage levels
than you get from a microphone (surely there is some trade off
here* ? Trafos for MC phono cartridges IME are tiny) then is it
not as a first question what is the likely best secondary load
given the source impedance is 2K maximum ?
Is there anything that means it simply will not work well ?
Are there any real bass issues ?
Are there any real treble issues?
rgds, /sreten.
* For MC cartridges I understand you need to make the transformer as
small as possible for the given maximum levels such that very small signals
are not swamped by the high magnetisation issues of an overlarge core.
Cool, OK. I've seen the ringing and yes, A Zobel can help.
Michael, can you elaborate further on the on voltage headroom and Faraday?
Most of my CDP's have output impedances around 1k, so that's what I work with.
My passive pre is a shunt type, with a 100k pot and 10k series resistance. With the primary impedance of 14.7k, reflected from the 367k secondary load, the CDP will see impedances from 10k to 22.8k, satisfying the 1:10 ratio at all times.
Cheers,
IG
According to laws of physics, of course!
Suppose, you have a voltage divider formed by a 150 Ohm output resistance of a microphone, and an inductance of 1H shunted by 150 Ohm load resistance. It is very rough scheme that reflects only primary inductance and reflected to the primary load resistance. You may calculate frequency response and find that it satisfies requirements.
For an optimally designed microphone transformer primary inductance is not higher than required for frequency response on low end, but according to given source and load resistances. Why? Because trying to improve LF responce beyond specs you affect HF response increasing parasitic parameters.
That means, if to use such a transformer with higher source and load resistances, you will get higher attenuation on low end, that will be out of specs. I even don't speak here about effect of core non-linearities, just about basic things. If you disagree with them that means you have to learn basics carefully.
This is my point.
Also, as I said initially, I don't have an experience with this particular transformer and don't know how optimally it was designed for 150 Ohm microphone input, so it may work satisfactory, but may not.
No, you're right on the mark, I would say. I think the confusion comes in when you fail to differentiate between power transformers and signal transformers.
The example that was brought up regarding increasing voltage at the expense of decreasing the current was using a 110V / 220V comparision, implying power transformers. With that specific example, the rationale falls apart. Any competently built (but also cost effective) power transformer optimizes the core size and volts/turn to run near core saturation at nominal ratings. So doubling the excitation magnitude (same frequency) will certainly saturate the core and drastically distort whatever transformation is left, regardless of current magnitude. Can't be done.
Signal transformers are a tad different. They are typically rated for impedance ratios, somewhat adding to the confusion for all. Yes, impedance ratios imply turns ratio, but in my mind the optimal spec would be turns ratio along with intended signal strength (or at least actual winding impedances). In some instances you can infer signal strength to impedance, but this does not guarantee you are operating the transformer in its optimal range. The 600 ohm transformer is most abused, with many insisting you should insert a 600 ohm source impedance and terminate with a 600 ohm resistor; a fallacy.
Therefore, with a signal transformer you are able to trade off voltage with current, as long as you don't enter saturation (which is a function of voltage and frequency, NOT current). So what is saturation of a 600:600 transformer ? Undefined. Need more information.
Hi,
I do not know but I will say its a signal transformer not a power transformer, you
cannot load the output of any electronics with anything like their nominal output
impedance, which you can with a dynamic microphone and get optimum power
transfer = lowest noise performance. So I cannot understand your numbers.
AFAICT the transformer is designed for 1.5k primary due to a 39k secondary.
rgds, sreten.
Yes, I believe that is correct. The mic channel input impedance of many mixing consoles is ~1.5K . That seems to fit the widest range of mics.
That many mixing consoles that you mean do not have input transformers. 😉
1:1:16.3 transformers that I use in my vacuum tube powered mixer have 3 Henry primary inductance when connected for 150 Ohm:40KOhm. And they are fair 150Ohm/40KOhm transformers, according to measurements. When I switch to 600 Ohm/40KOhm ratio I measure 12 Henry primary inductance.
I can't reveal the source because I do not want to skyrocket prices on them.
The truth is, most of typical mainstream mixers have so called "Electronically Balanced" inputs. Read, SMD opamps.
How much would cost transformers for a typical mixer with 16 input strips? How much that typical mainstream mixers cost? Mic input transformers in my 8 channel mixer cost more than that mainstream 16 channel mixers.
- Status
- Not open for further replies.