Looking for advice for a non-Bottlehead product

[PassionForSound]

I'm about to build and review a kit amp that's similar to the S.E.X. in the sense that it is SET and drives both speakers and headphones. However, the problem with it is that the output impedance for headphones is a bit ridiculous (~200 ohms!)

Based on the fact that the S.E.X. performs wonderfully doing much the same job, I can't fathom why they would have placed this level of impedance in the headphone output path. Can any of the gurus here explain to me if there are risks (other than high output volume) in bypassing or at least reducing the value of the resistors in the HP output path?

And to help me provide a good solid review, how is this same challenge managed in the S.E.X. without creating stupid levels of output impedance?

 

[Paul Birkeland]

So the amp would otherwise have very low output impedance but there are 200 ohm resistors between the speaker output and the headphone output?

The ~2006 edition SEX amp (2.0) had resistors between the speaker output and this is PJ's commentary on the subject:

"The IHF standard calls for a 5vRMS source with 120 ohms impedance. This is supposed to be applicable to all headphones. It will produce maximum power into a 120 ohms headphone; that power is volts squared over resistance, in this case 0.052 watts. At the extremes of 600 ohm phones, or 24 ohm phones the available power is less, 0.029 watt. So for a tube amp capable of delivering 0.05 watts (50 milliwatt or mW) an output transformer with taps for the various headphone impedances in the range of 24 to 600 ohms would be appropriate, in order to preserve an appropriate load on the tube to keep distortion low.

As the available amp power increases, transformers with lower impedance outputs will be able to supply the necessary voltage. Used in this way, the amplifier always sees a high impedance load and distortion is kept low. At the same time, the noise voltage is reduced by the lower impedance output tap. For more information on this issue, look at the Bottlehead Community page; my white paper on signal and noise levels is near the top. The SEX amp produces 2 watts, which is 4vRMS, on the 8 ohm transformer tap loaded with a speaker. With the standard 120 ohm series resistor it can produce 5 vRMS easily, thus meeting the IHF standard accurately and with a very low noise floor.

There is a 500 ohm tap available on the stock autoformer. However, it would produce over 30 volts into a set of headphones, far too much for the phones or your ears to handle.

Now admittedly the IHF standard is old, and specifically it predates the widespread use of headphones with portable devices like iPods. These devices run on low voltages, and in order to obtain sufficient loudness the headphones need to be low impedance and highly sensitive. Some of these headphones do not work at their best with the IHF standard 120 ohm source impedance, being designed for a near zero source impedance and low voltages (around 1vRMS is not uncommon). Unfortunately, few headphone manufacturers will tell you this. But if you do have such a set of phones, the ideal would be a low-impedance source of even lower voltage - in other words, for the SEX amp a tap of less than 1.0 ohms, with no series resistor. It is easy enough to modify the 120 ohm series resistor to an L-pad which will reduce the output voltage and impedance for such phones, though I don't think we have documented such a modification yet. It would merely involve changing the 120 ohm resistor for another value, and adding another resistor from the "hot" headphone signal to ground. I would suggest 6.8 ohms (5 watt) and 1.2 ohms (1 watt) as widely available and suitable starting values."

For the newer SEX amp with the current hot headphones, the resistors no longer made sense and instead we just worked on making it super quiet so it would work well without the resistors into highly sensitive headphones so that maximum power could be available for those newer power hungry headphones.

 

[PassionForSound]

My sincere thanks to you Paul and to PJ for the post you quoted. This is exactly what I needed!!

I'll look at adding an L-pad to modify the output with in order to manage output noise and still keep the OI a bit lower.

 

[PassionForSound]

Hi all, I've got this kit now and it's all built and sounding good (wish I still had a S.E.X. to compare), but the output impedance is an issue and the setup.theyve used to tame the output has me confused.

I'll attach the circuit diagram and would dearly love someone to confirm my thinking. Please forgive any incorrect terminology. Hopefully I make sense despite potentially using the wrong words sometimes.
Diagram here: https://photos.app.goo.gl/K9Z1AuYMzbFJvkK4A

There are three output levels selectable via a jumper. All of the following is based on the low output setting.

It seems to me that the signal passes from the output into the headphone, but with a connection siphoning some signal off by resistors. It appears to me that in low mode, there is 9.9 ohms attached off the main signal path (3 X 3.3 ohm resistors).

After the headphone, on the return path of the signal to ground, there is a 220 ohm resistor then the connection of the 9.9 ohm 'siphon' resistors and finally an additional 3.3 ohm resistor.

Is how I'm reading this correct? If so, what would be the purpose of the 220 ohm resistor instead of a simple L-Pad type setup with lower values?
Finally, any recommendations for fixing this to keep a reduced output voltage while also dropping the output impedance to something more manageable (10-15 ohm would be ideal, but anything significantly below 220 ohm will be welcome)

 

[Paul Joppa]

The resolution of the image is too low to be legible.

 

[Doc B]

Have you contacted the manufacturer to see what their take on this is?

 

[PassionForSound]

The resolution of the image is too low to be legible.

If you click on the image after opening, you can zoom in and out will resolve after a second or so with higher quality, but I'll also see if I can get a higher quality pic attached ASAP

 

[PassionForSound]

Have you contacted the manufacturer to see what their take on this is?

Sadly, they're very rigid in their response - words to the effect of "the designer will not alter his design for a handful of DIYers"

It's a bit ifvav disappointing response, particularly in light of this community

 

[Paul Birkeland]

Yes, that did resolve.  So if you have a power amp with a lot of power and a lot of gain and you want to use headphones on it, there are a couple of hurdles to deal with.  The first one is to determine whether the amp will have issues if it's run without a load.  If yes, then you need to deal with loading the output.  This is part of what's going on with the series strings of 3 ohm resistors.  In the absence of a speaker, that still provides a load at the output.

Another issue you can have is when trying to make an amp that is both quiet enough for something like an IEM but at the same time has enough power to run a big planar headphone.  This is the other portion of why there are a bunch of 3 ohm resistors in series.  You can pad the output down substantially in the low output mode, then the amp's noise floor is also knocked down at the same time, but on the flip side you have the option of much greater power/voltage on the high output setting. 

As far as the 220 ohm resistors go, PJ's commentary on SEX 2.0 explains why we did it, and you seem to have a very firm grasp on why we don't do it anymore!

 

[PassionForSound]

Thanks Paul. Based on this design, is there any issue you can see with me replacing the 220 ohm resistors with a lower impedance L-Pad and leaving everything else intact?

 

[Paul Birkeland]

That would be a question you'd have to ask the manufacturer.

 

[Paul Joppa]

If you click on the image after opening, you can zoom in and out will resolve after a second or so with higher quality, but I'll also see if I can get a higher quality pic attached ASAP

Thanks - I made the mistake of downloading the image, which gave me a smaller file. Now I can read it easily.

Yes, PB has accurately summarized the issue and you seem to have correctly understood it. The four 3.3-ohm resistors are tapped at full voltage, -6dB, and -12dB, after which the 220 ohm resistor is always in series with the phones.

In light of the manufacturer's response, I would not modify the amp (they can get really cranky about such things). For low impedance phones, you might externally put a resistor in parallel with each headphone driver, thus using the 220 ohm resistance as the series arm of an L-pad. If you do that in a separate box, you could switch it in or out to hear the difference.

If the resistance is equal to the headphone driver impedance, that would reduce the level by 6dB and give a unity damping factor; if it's 1/3 the driver impedance you'd get -12dB and a damping factor of 3.

 

[PassionForSound]

That would be a question you'd have to ask the manufacturer.

Fair enough - thank you for your help to this point.

Thanks - I made the mistake of downloading the image, which gave me a smaller file. Now I can read it easily.

Yes, PB has accurately summarized the issue and you seem to have correctly understood it. The four 3.3-ohm resistors are tapped at full voltage, -6dB, and -12dB, after which the 220 ohm resistor is always in series with the phones.

In light of the manufacturer's response, I would not modify the amp (they can get really cranky about such things). For low impedance phones, you might externally put a resistor in parallel with each headphone driver, thus using the 220 ohm resistance as the series arm of an L-pad. If you do that in a separate box, you could switch it in or out to hear the difference.

If the resistance is equal to the headphone driver impedance, that would reduce the level by 6dB and give a unity damping factor; if it's 1/3 the driver impedance you'd get -12dB and a damping factor of 3.

Thanks for the assistance. I ended up changing tack today and creating an external box that connects to the speaker taps and allows me two different output levels (22 ohm & 1 ohm L-Pad and 39 ohm & 1 ohm L-Pad). So far things are sounding pretty good!

One question though, I've noticed that I can hear music from the amp with nothing plugged in and with the volume at about 50% or more (the amp is back in it's stock form so no weird mods involved). I'm assuming this is a phenomena with the tubes and I've either never noticed it before or the Crack, SEX and Mainline don't do it. Am I right that this is a normal thing with some tube amps?

 

[Paul Joppa]

I think the audible sound with nothing connected to the  outputs is more likely to be from the output transformer. Both electromagnetic forces and dimensional magnetostriction have roles in transformer vibration. I don't know enough about it to speculate why it may be greater or smaller in particular designs.

 

[Doc B]

Yeah, it's probably a singing output transformer. I've come across this a few times, usually when I was building clip leaded prototypes way back in the day.

 

[PassionForSound]

Thanks guys. Sounds like you're not overly concerned which gives me comfort. I'll likely be upgrading the OTs in the future so it'll be interesting (for me) to see if the singing is due to the circuit design or just the specific OTs.