You could have googled 111C and found the actual connection data.
Correct.
No. Ignore the fact that the amp uses one termal as "ground". Connect the red to one input terminal and the black to the other.
Here's a few measurements of a 111C. It was driven from a power amp capable of 48Vrms/+35.8dBu, through a 600 ohm build-out resistor, and the secondary was terminated by a 600 ohm resistor.
First, frequency response measured to 50KHz (phase shown in blue):
As you can see, it's down .3dB at 20KHz, .6dB at 30KHz. In other words, no audible "rounding" off of high frequency response.
Now for some distortion. This plot was made at just a tad under +36 (as applied to the total system including resistors):
The only audible "color" (distortion) occurs below 40Hz at this rather high level, and frankly, it remains in the questionably audible level until around 20Hz. The rest of the spectrum is remarkably clean.
Lastly, stepped distortion vs frequency. Each run was made by increasing the drive 2dB. Overall distortion seems to go down with increasing level because this system looks at THD+N (noise), so actual THD doesn't change much, but the signal to noise ratio increases as the drive signal goes up. Also, the "0.0" plot in this set is at +35.8dBu, so the distortion max loosk a little higher at 20Hz. However, it remains in the inaudible range for the rest of the spectrum. Also, the distortion vs frequency curves are somewhat different in this measurement system due mostly do different FFT parameters. I didn't take the time to match the systems to each other this time, sorry.
While it is possible to measure different source Z and load Z combinations, given these results it didn't seem worth the effort. I did try one distortion set with no build-out resistor (driving from very low Z) and a bridging load. Distortion at 20Hz went down slightly, not remarkably. Unfortunately I mistakenly failed to save the data.
For reference, the graphs with the black background were made with REW using the QA401 interface. The white background distortion series is the QA401 system and software. In this graph the QA401 output is shown directly in dBV, but there is a large 1000W power amp between it and the 111C, along with the built-in pad, plus an additional 10dB pad to protect the interface from damage from over-voltage (probably unnecessary but I hate blowing stuff up accidentally). In the full test system, QA401 0dBV = +35.8dBu/48Vrms.
What we have here is a transformer that is basically incapable if providing any classic "transformer color". It's not the only one that is this good, but there are plenty of transformers that create massive signal modifications. Just not a 111C. It was darn good 60-70 years ago, and still is. But colorful, nope.
First, frequency response measured to 50KHz (phase shown in blue):
As you can see, it's down .3dB at 20KHz, .6dB at 30KHz. In other words, no audible "rounding" off of high frequency response.
Now for some distortion. This plot was made at just a tad under +36 (as applied to the total system including resistors):
The only audible "color" (distortion) occurs below 40Hz at this rather high level, and frankly, it remains in the questionably audible level until around 20Hz. The rest of the spectrum is remarkably clean.
Lastly, stepped distortion vs frequency. Each run was made by increasing the drive 2dB. Overall distortion seems to go down with increasing level because this system looks at THD+N (noise), so actual THD doesn't change much, but the signal to noise ratio increases as the drive signal goes up. Also, the "0.0" plot in this set is at +35.8dBu, so the distortion max loosk a little higher at 20Hz. However, it remains in the inaudible range for the rest of the spectrum. Also, the distortion vs frequency curves are somewhat different in this measurement system due mostly do different FFT parameters. I didn't take the time to match the systems to each other this time, sorry.
While it is possible to measure different source Z and load Z combinations, given these results it didn't seem worth the effort. I did try one distortion set with no build-out resistor (driving from very low Z) and a bridging load. Distortion at 20Hz went down slightly, not remarkably. Unfortunately I mistakenly failed to save the data.
For reference, the graphs with the black background were made with REW using the QA401 interface. The white background distortion series is the QA401 system and software. In this graph the QA401 output is shown directly in dBV, but there is a large 1000W power amp between it and the 111C, along with the built-in pad, plus an additional 10dB pad to protect the interface from damage from over-voltage (probably unnecessary but I hate blowing stuff up accidentally). In the full test system, QA401 0dBV = +35.8dBu/48Vrms.
What we have here is a transformer that is basically incapable if providing any classic "transformer color". It's not the only one that is this good, but there are plenty of transformers that create massive signal modifications. Just not a 111C. It was darn good 60-70 years ago, and still is. But colorful, nope.
With the same voltage level at the amplifier output or the same voltage level at the transformer input?
With 35.8 dBu at the amplifier output and with 600 ohm between amplifier and transformer and 600 ohm load, you are essentially driving it with its specified level of 30 dBu. It certainly can handle that quite well!
Amen to that.
Very cool and impressive. Although you did show in one of the graphs that there is a high frequency rolloff of the 111C that appears to begin at 5k. However it doesn't sound like that is a function of level but rather the natural frequency response of the unit. It's down 10db by the time it gets to 20k. That would no doubt sound like you put a high shelf filter on the signal, no? Which also could provide a "rounding" affect in my opinion. I wonder if that's what people hear.
I want to run some LUFS numbers with music when I get my unit set back up. Most interested in peak to loudness and range measurements. I do know that in the original files I null tested, the dynamics of the 111C version were slightly lower, which would indicate that there is some truncation of peaks.
Last edited:
Hey. So I have to rewire my 111C to match the first schematic you drew, but so far I only got the 560ohm resistors that jaddie originally suggested, that will go between the amp and the 111C. Do you know how much max level that will end up putting into the 111C? I tried to do the calculations myself and got something like 26dbu, but that is not considering whatever happens between the 111c and the interface it goes back into. I.e. what the effect of the -30db pad and the 10k ohm input resistance value will do. I'm talking about without any of the resistors on the back half of the circuit. Just the 560s on the front.
Regarding post #166: about 22 dBu on the secondary side, neglecting transformer losses and the fact that the amplifier will produce a bit more voltage at light load.
With the primaries in parallel and the secondaries in series, you have a 1:2 transformer. The impedance goes with the square of the turns ratio (because when you transform the voltage up, you transform the current down and the other way around). Hence, the 1010 ohm of the loaded attenuator ideally becomes 252.5 ohm on the primary side.
The RMS voltage at the primary is then the RMS voltage coming out of the amplifier, sqrt(4 ohm*60 W), times 252.5 ohm/(252.5 ohm + 560 ohm); that is the standard voltage divider equation. It's 4.8144 V.
You get twice that at the secondary, so 9.6288 V.
Expressed in dBu, that becomes +21.89 dBu.
With the primaries in parallel and the secondaries in series, you have a 1:2 transformer. The impedance goes with the square of the turns ratio (because when you transform the voltage up, you transform the current down and the other way around). Hence, the 1010 ohm of the loaded attenuator ideally becomes 252.5 ohm on the primary side.
The RMS voltage at the primary is then the RMS voltage coming out of the amplifier, sqrt(4 ohm*60 W), times 252.5 ohm/(252.5 ohm + 560 ohm); that is the standard voltage divider equation. It's 4.8144 V.
You get twice that at the secondary, so 9.6288 V.
Expressed in dBu, that becomes +21.89 dBu.
So wired exactly as you have it, I will only get +21.89?? Lol. That's less than what I am currently sending to it with the SSL! Sigh. I thought the power amp solution was gonna be easy. Guess not. This is tempting me to just go out and find a more powerful amp and resell the Samson when it gets here.
**And just to add, the wattage rating on the amp is almost certainly peak values, not rms. And as jaddie has said, we are really looking at peak values going into the 111c. Does your 21.89dbu value increase if you're looking at it in terms of peak values??
Yes, with a too high resistance on the primary side and a too low resistance attenuator on the secondary side.
It is almost certainly neither. No-one in their right mind calculates the root of the mean of the square of an instantaneous power.
I'm just thinking back to my car audio days. Most of the wattage ratings were peak values. They put those up instead of RMS so they could get away with putting a higher wattage value on the label. Maybe I'll email them and ask which it is.
Well I guess I better get the resistor parts in your schematic so I can try it out after the original plan proves to be too weak.
No I didn’t. Look again.
Not at all. I posted the value for 20KHz. Did you read what I wrote even if you can’t read the graph?
It’s nothing like a shelf response. It’s a roll off response, just no in the audio band. I shelf is an initial rolloff then a flattening. Wow we have some work to do here.
Argh. There’s nothing in the graph that is audible to anyone.
No, LUFS won’t show you any difference at all, and would be completely dependent on settings elsewhere in the system, including power amp input gain.
Dynamics cannot be changed by a linear device.
Response scale is on the left, phase scale is at the right. I’m so sorry I posted the phase plot, it confused you.
You didn’t understand what the car guys were doing either. But it’s out of context here.
How does this part list look?
Ceramic Composition Resistors 2W 150 ohm 10%
HPC2C151K
KOA Speer
Thin Film Resistors - SMD 3.3K OHM .1% 10PPM 1/10W
RT0603BRB073K3L
YAGEO
Thin Film Resistors - SMD (680 ohm)
RT0603CRE07680RL
YAGEO
So.. apparently you either can’t read the graphs I posted, or don’t believe them, or both.
Let me be clear here. I ran the tests with your schematic with a 1Kw amplifier and couldn’t get any high end roll off or any significant distortion out of it. Get all the parts you want, nothing here is going to change. The change you’re looking for, high end “rounding” (not a term used in audio) can be achieved with any EQ plugin far better, and under your complete control. It’s a 1st order low pass response. Put the -3dB point at 20KHz to start. Not a shelf. You can also do this OTB with a resistor or two and a capacitor. And you can rid yourself of digital distortion by keeping peaks below 0dBFS.
All along the way I’ve tried to help you avoid wasting time and money. I’ve given you scientific proof that the 111C is the wrong device to do what you’re asking, but clearly that’s what you want to do. I’m sorry, I can’t help any further.
There's not really a lot to understand. It's well known in the car audio industry that a lot of manufacturers post only their peak watt ratings to make people think they're more powerful than they really are. That's the only point I was making, that because the Samson didn't say whether the watt rating was RMS or peak I was wondering if it could in fact be peak. Maybe the power amp industry doesn't do the same thing that some people in the car audio industry does. And if the ratings on the Samson are rms, then one could assume that it's capable of a higher wattage for peaks. That doesn't seem like an unreasonable question.