Trying to understand tranformer loading....

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Sender

Well-known member
Joined
May 17, 2005
Messages
242
Location
USA
So I've been doing alot of reading about transformer loading and have a couple of questions. I guess I should first state the transformer and circuit I'm going to use. It's a Jensen JT-115K-E (http://www.jensen-transformers.com/datashts/115ke.pdf) into an API 312.

So to my understanding, I shouldn't have to add aditional load to the front end of the transformer because it's designed to handle microphone impeadance. However the output of the transformer does need to be loaded to get the correct impeadance into the 2520.

Without load, the 115, has an output imeadance of 17k. In all the 'typical application' circuit on the data sheet, they have a 125k resistor in paralel on the secondary to give a total output impeadance of 83.5k (Rt = (R1+R2)/2)[R1 being 17k (the output impeadance) and R2 being the 125k resistor).

So I've replaced R1 ((http://www.danalexanderaudio.com/ApiInfo/api325schem.jpg) I'm using a 325 retrofitting to a 312) with a 100K ohms.

The 2520 has an input impeadance of "greater the 10M ohms". So let so 10M ohms is the minimum. I read somewhere that briding (having at least 10 times the destination impeadance then the source impeadance) is ideal to transfer the maximum amount of voltage and to minimum amount of distortion and frewuency chage. So the 10M ohm input impeadance is around 120 times the source impeadance of 83.5. Is this okay?


Am I correct in doing so? Or is my logic all flawed?
 
Hi sender.

The 2520 is a BJT input device and noise would not be optimal with a 10M R on its input.

125K is fine.
 
The Jensen data sheet should provide a good termination resistance value to keep the transformer happy.

The bridging rule of thumb (10x source impedance) is most important at the mic interface. You typically want to keep loading at the mic input on the order of 10x the microphones 150-200 ohm source impedance.

The effective impedance at the input of the transformer will be the output load reflected back and reduced by the square of the turns ratio. So if the transformer turns ratio is X, the impedance will be reduced by x^2.

The 115KE looks like 10x turns ratio so secondary load will reflect back divided by 100. Jensen's recommended 150k termination will provide approx 1.5k at input. I wouldn't go much larger than this in case the transformer is optimized for frequency resonse at that loading, and much less than this will load the microphone too heavily and hurt voltage transfer.

JR
 
[quote author="Kit"]Hi sender.

The 2520 is a BJT input device and noise would not be optimal with a 10M R on its input.

125K is fine.[/quote]

The impedance at the secondary of the transformer will be the termination in parallel with the mic's source impedance. For the 10x transformer this should be dominated by the mic and will be in the 15-20k ballpark. This is perhaps borderline high but usable with a good BJT OA.

JR
 
Yeah, use a little intuitive guess work, try a value, then work up or down from there.
Every mic will want a slightly different resistor there, so it's always a compromise, unless you are going to "JB Weld" your signal path together, for the next 100 years, which, BTW, was probably quite a common occurance back then, the same mic, at the TV station, running into the same board for years and years, never gets moved, well, I do not think that is your case, so ge it as close as you can, and then just eq any high end tweaks that you need when changing mics/rooms/git players.
You could always put in a rotary, but the input of a high impedance amp is not a good place to run wires from.


The resistor deals only with the high end.
At the low end, the load from the xfmr shunts everything out, due to droped inductance/impedance from the 2 pi F L formula.
 
Firstly, what does BJT stand for?

Secondly, I was noticing that output impedance of a SM57 is 300 ohms, so with the 150K on the secondary, it only provides times 5 (approx) difference. Is this enough for bridging??

----------------------------

Onto a similar note, my circuit isn't working....

The way I have it running right now is that red is pin 2 on my xlr, brown is pin 3, then yellow goes to hi in (pin 5), orange to lo in (pin 4), then i have a 125 ohm resistor across pins 4 & 5, then pin 4 is tied to pin 3 (com). When ever I have a mic connected to it, the circuit buzzes, passes no audio, and as soon as i remove the mic, no buzz.

Does anyone have any recomendations?

Sorry guys, i'm a bit of a noob! I really want to learn though. I apprieciate all the help!!!
 
What John Roberts said.

Essentially, a transformer works optimally with a particular set of source and load resistances. In the case of this particular Jensen that would be a 150-200 ohm source resistance (the microphone) and 150k load resistance (the terminating resistor). This transformer was designed to be used without a Zobel network on the output.

As John sez, the 150k resistor reflects back through the tranny to be 1500 ohms. Well, approximately; the secondary DC resistance of 2960 ohms adds to the load resistor, so it's really about 153k, and the reflected load is now about 1.53k, plus the primary resistance of 28.7 ohms, for a total of 1558 ohms. In parallel with the phantom power resistors that will be about 1398 ohms at the input. That's what the mic sees, and that's good for most modern microphones.

What's at the output? Looking back, you've got a 150 ohm source (typically), in parallel with the phantom resistors for a total of about 148 ohms, plus the tranny's series resistance of 28.7 ohms for a total of about 177 ohms. That's stepped up by the transformer to about 17.7k plus the secondary DC resistance, for a total of 20.7k; the loading resistor is in parallel with that, so you wind up with about 18.2k. This is the impedance your opamp sees at the + input. Add to that the parallel resistances of the resistor going from the - input to ground, and the feedback resistor, and you get the total source impedance the opamp is seeing.

Whether that will produce a reasonable noise figure on a bipolar transistor opamp depends on what the input transistors are and how they're biased. If you take the opamp's e(n) and divide it by the i(n), you'll get the optimum source resistance for that particular opamp. The closer you are to that, the less noise the electronics will add.

As I mentioned, the input Z of about 1.4k will be good for most modern microphones. In my experience, mics in the Shure SM57 family will prefer a lower load impedance; add a resistor in parallel with the input, about 787 ohms. Old-style ribbon mics (RCA, AEA, etc.) will be happier with a higher load impedance, which won't be easy to do without compromising the transformer's frequency and transient responses.

Peace,
Paul
 
[quote author="Sender"]Firstly, what does BJT stand for?[/quote]

Bipolar junction transistor. As opposed to a field-effect transistor.

Secondly, I was noticing that output impedance of a SM57 is 300 ohms, so with the 150K on the secondary, it only provides times 5 (approx) difference. Is this enough for bridging??

Yes; in fact, my experience has been that SM57s actually prefer a lower load impedance, someplace between 500-600 ohms. They sound less harsh that way.

Peace,
Paul
 
Production gets ready to start shipping a new design, the QC Dept guy has a beef with the be-spectlacled engineer because he's screwin his wife, red tags the xfmr because the graph on the data sheet is ten miles from the graph on the QA scope, they can both smell the UPS truck next door, so they both go outside for a smoke to discuss the situation, (actually back then, you just light up inside. remember those days?)

Engineer says, "Look, I'll spend 1/2 an hour with the decade box, but thats it."

Resistor gets added, Drafting Dept. records the wrong value if at all, field tech finds out that the xfmr runs better un loaded because of all the cabling so he rips out resistor, DIY guy takes apart board 30 years later, puts schematic on the web, engineer reads forum....

need I go on?
 
[quote author="CJ"]... (actually back then, you just light up inside. remember those days?)
[/quote]

A story from an older engineer where I used to work... When engineering was all-tube (read: tubes put out lots o' heat) there was a guy who was often discovered naked, smoking a cig, and finishing up a 6-pack by the time the mid-morning shift got in. This was in the early 70s from what I understand.

This thread hasn't covered loading + inductance vs. frequency response, which I'm working on a methodology for practical usage thereof. Any additional technical comments concerning inductance+loading/freq. would be great too!
 
CJ, I remember! I worked at IBM circa '85 and you could smoke just about everywhere except the clean rooms (for obvious reasons!). Later, they banned smoking in most of the common areas but set up well-ventilated smokers' lounges in several locations throughout the plant. That was a good compromise that seemed to make everybody happy.

One of my other long-ago jobs was stuffing circuit boards at another company. They let us smoke right at our workbenches, although we had to hide the ashtrays if the client came around for an inspection :roll:. Even as recently as the mid '90s, I worked in an MI repair shop where I could smoke at my bench. Of course, now you risk a fine and a horsewhipping if you're seen lighting up anywhere outside your own home or car.

Back on topic: I've heard others suggest, and have confirmed this for myself by careful listening, that the SM57 seems to sound less harsh when working into a transformer-coupled input. It does tend to sound strident into a bridging, transformerless input (with a mostly resistive input impedance), so with that sort of input I'd definitely experiment with shunting as per Paul's message. But you might shunting unnecessary--or at least less necessary--with a transformer-coupled input. Transformers and preamp circuits differ so it's hard to generalize; just experiment and listen.
 
I would advise against dialing in your mic preamp for optimal sound with a SM-58, or any one mic for that matter unless you live on a desert island and only own that one mic. The dominant change due to termination is mainly frequency response, which is why they invented EQ. Get the voltage with good S/N first and deal with EQ later.

The SM-58 is so popular because of it's ruggedness and repeatability, not because of it's special sound. Especially for use in live sound reinforcement, it's important for the mic to have a well defined pattern, sensitivity, and a consistent response curve so you don't get nasty surprises like your monitor system feeding back, etc.

I am suspicious that some high end mic preamps intentionally go low, or high with their input termination so they will sound different in comparison to normal (read flat) preamps. This will not show up in typical frequency response test performed with a low resistive source impedance signal generator, but you could hear the difference with some real mics. You can't expect people to pay more money if the preamp sounds the exactly the same as the cheaper models. :roll:

JR
 
[quote author="NewYorkDave"]I've heard others suggest, and have confirmed this for myself by careful listening, that the SM57 seems to sound less harsh when working into a transformer-coupled input. It does tend to sound strident into a bridging, transformerless input (with a mostly resistive input impedance), so with that sort of input I'd definitely experiment with shunting as per Paul's message. But you might shunting unnecessary--or at least less necessary--with a transformer-coupled input. Transformers and preamp circuits differ so it's hard to generalize; just experiment and listen.[/quote]

Indeed, experiment and listen. We tried the SM57 into a Universal Audio 2-610 preamp, with one channel set for 2k input impedance and the other for 500 ohms. Big difference. Both channels were transformer-coupled, with the lower-Z coming from a tap on the primary.

So we tried using a resistive shunt on the second channel instead of the tapped primary. It had essentially the same effect on the SM57 as going into the tapped primary: less glare, less audible distortion, smoother sound. That was a transformer shunted by a resistor.

By the way, so far the SM57 is the only mic we've tried that showed a major improvement. Since the SM58 and SM56 use the same capsule I'd expect similar results with them.

Peace,
Paul
 
The 57's I have were modified by removing the internal transformer so the voice coil is connected directly to the mic output. Since I only use these on snare drums and very LOUD guitar amps, they have plenty of output and no usual 57 crunch. Using the impedance switch seems to only affect level of these mics, and not response. It's an old trick lots of PA guys did with their 57's.
 
Thanks for the help guys!! I started this thread with hopes of getting a mic pre to work that wasn't working (I was hoping that improper loading was the issue). I think I did it right, but let me double check.

I have the input of the transformers wired to pins 2 & 3 on my XLR (what should I tie pin 1 to, 0V? Chasis?). Then I have the secondary tied to pins 4 and 5 (lo in and hi in) of the edge connector. Then I have a 125k resistor, well two resistors that add up to 125k, across pins 4 and 5 (which i'm under the impression that this is the correct way to load the transformer). Then I have pin 4 and pin 3 (0V), per recomendations in the various API threads.

So this is a correct configuration for tranformer loading, correct?

So if this is correct I need to test my circuit and figure out where the signal stops. If this was line level, it would be cake, just hook up my tone generator and grab the oscope and figure where it siezes. However, I don't have a mic level generator.

So secondly, how can I get line level to mic level? I was thinking if i put a resistor across pins 2 and 3 of my generator, i'd be able to reduce the impedance to only as low as 300 ohms, not 150. So can i achieve mic impedance from line level for testing?
 
To use the line-level output of your signal generator to test a mic preamp, build a 40dB L- or U-pad: a 15K series and 150-ohm shunt resistor (unbalanced) or two 7.5K series resistors and one 150-ohm shunt resistor (balanced).
 
also, you asked about where to connect pin 1 from your input... since it's a mic preamp, and you likely want a phantom supply (an assumption) you'll need to tie that pin1 to 0v so that 48v will have a reference.

nice tip on the pad dave, i think i'm going to build up a cable for it actually, bnc to xlr male, and put the resistors in the xlr for handiness.
 
[quote author="Sender"]I have the input of the transformers wired to pins 2 & 3 on my XLR (what should I tie pin 1 to, 0V? Chasis?). Then I have the secondary tied to pins 4 and 5 (lo in and hi in) of the edge connector. Then I have a 125k resistor, well two resistors that add up to 125k, across pins 4 and 5 (which i'm under the impression that this is the correct way to load the transformer). Then I have pin 4 and pin 3 (0V), per recomendations in the various API threads.

So this is a correct configuration for tranformer loading, correct?[/quote]

Actually this transformer wants to see a load of 150k, not 125k.

Peace,
Paul
 
Back
Top