How dangerous are output transformerless tube headphone amplifiers really?

[rock soderstrom]

Hi folks,

I've been using a self-made OTL headphone amplifier for over 12 years without any problems. The design comes from John Broskie, it is a 6CG7 / E88CC Aikido amp, essentially a gain stage with an active load and a optimized White Cathode Follower circuit.

This amplifier sounds really very good, I also use it as a final monitoring amplifier for my productions.

I have also built HPAs with transformers, unfortunately the really good ones are very expensive and rare.

Now I would like to build another OTL HPA as a Christmas present for a friend and wonder how dangerous are these ironless amps really? It's one thing when I fry myself and quite another when my buddy (or his kids) are getting harmed.

Do not get me wrong, I can safely set up such devices in general, I have enjoyed the appropriate training. I'm only talking about the possibility that the last coupling capacitor will let the operating voltage pass through to the headphones in the event of a fault.

How likely do you think that is? What would really happen? What additional security measures could one establish? How dangerous are OTL HPAs in general?

B+ is 250V for the planned amp, the output driver is biased around 20mA per channel. (ECC99). The amp is designed for 250-300R headphones.

Cheers!

 

[emrr]

Aren't these always cathode followers?  Meaning low voltage on the output side.  There are certainly commercial versions on the market, which have passed safety certifications.

 

[rock soderstrom]

Aren't these always cathode followers?  Meaning low voltage on the output side.  There are certainly commercial versions on the market, which have passed safety certifications.

Thanks for your input. Low voltage is relative. The normal operating voltage on the HT side of the output cap is B+/2.
In a disaster scenario, it could also be full B+, which is why I dimension the output capacitors to more than B+.

I am not sure if one could pass the safety certifications with a comercial product like this on my side of the pond.

I spoke to a few people from this industry whose opinion was very cautious to negative. OTL designs seem to be valued differently on both sides of the Atlantic? We tend to overregulate a bit here. German Angst?

 

[emrr]

Ah. 

My only initial thought, which won't be attractive, is series output caps for double isolation.  One RCA broadcast console did that with cathode follower preamps to reduce DC bleed and channel switching clicks. 

 

[abbey road d enfer]

Do not get me wrong, I can safely set up such devices in general, I have enjoyed the appropriate training. I'm only talking about the possibility that the last coupling capacitor will let the operating voltage through to the headphones in the event of a fault.

How likely do you think that is? What would really happen? What additional security measures could one establish? How dangerous are OTL HPAs in general?

Why don't you limit the output voltage with back-to-back zeners?

 

[rock soderstrom]

Thats an idea Doug. I have seen small circuits based on back to back z-diodes. I have no idea how good or bad this is??

Edit: abbey was quicker. Any bad effects on sound with these z diodes? More distortion?

 

[emrr]

Zeners crossed my mind, but I didn't want to bring that up without looking more closely. 

 

[abbey road d enfer]

Thats an idea Doug. I have seen small circuits based on back to back z-diodes. I have no idea how good or bad this is??

Edit: abbey was quicker. Any bad effects on sound with these z diodes? More distortion?

No added distortion until clipping.

 

[JohnRoberts]

No added distortion until clipping.

I don't know the impedances but zeners are higher capacitance. You can make lower capacitance clamps mixing small signal diodes around the zeners. May not be a concern for headphone impedances.

JR

 

[rock soderstrom]

No added distortion until clipping.

Thanks! I will test the back to back zeners next week and simulate the failure case with a test setup.

 

[rock soderstrom]

I don't know the impedances but zeners are higher capacitance. You can make lower capacitance clamps mixing small signal diodes around the zeners. May not be a concern for headphone impedances.

JR

Thanks, interesting!

Edit: very similiar schematic attached, I do not use the cathode cap.

 

[rock soderstrom]

My test rig, a self etched PCB with 12BH7/ECC99 loaded atm.

 

[CurtZHP]

I built one not that long ago.  If I managed not to kill myself, your friend should be fine.  ;D

 

[rock soderstrom]

I built one not that long ago.  If I managed not to kill myself, your friend should be fine.  ;D

Hi Curt, I'm glad you're still alive! ;D How do you like your amp? Which tube combination did you use?

 

[buildafriend]

Ah. 

My only initial thought, which won't be attractive, is series output caps for double isolation.  One RCA broadcast console did that with cathode follower preamps to reduce DC bleed and channel switching clicks.

simple as that? cool

 

[CurtZHP]

Hi Curt, I'm glad you're still alive! ;D How do you like your amp? Which tube combination did you use?

12AX7 for the input stages, 12BH7's for the outputs.
Did an article in "Radio World" about it...

https://www.radioworld.com/tech-and-gear/products/my-vacuum-tube-headphone-amp-project

 

[emrr]

https://groupdiy.com/index.php?topic=75029.msg951149#msg951149

 

[rock soderstrom]

Ahh, thanks! 12AX7 gainstage into a 12BH7 SRPP is a lot of gain!

My circuit got only roughly 20dB. It depends on the tubes in the first stage. Gain is ~ mu/2. The Aikido WCF is without gain as all cathode followers.

 

[Dreams]

Zeners crossed my mind, but I didn't want to bring that up without looking more closely.

I like the use of the phrase "crossed my mind" there

Here's a picture I enjoy from the Audio Cyclopedia. "Not Recommended"

 

[rock soderstrom]

I like the use of the phrase "crossed my mind" there

Here's a picture I enjoy from the Audio Cyclopedia. "Not Recommended"

Thanks for your input.
But the situation in your photo is different. The coil of the headphones serves as an anode load and sees constantly full B+ operating voltage.
Defnitley not recommended.

In a worst case scenario (capacitor goes 0 ohm) with an OTL amp, the applied DC voltage would be short-circuited to ground by the loudspeaker coil. The headphones would say goodbye with a loud "plob" as the coil would burn out immediately. Right?

I am not an expert, but I believe that with a Zener protection circuit you can get this HPA sufficiently safe. This is my current status, but I am still happy to receive any opinion. This is what this thread is for.