Remember These?

[Cressy Snr]

You could try taking the feedback from before the output cap and adding a blocking cap in the feedback loop (can be much smaller). Stops the feedback from working to correct the output cap low pass filter.

That sounds like a good idea to me

[IslandPink]

Ah yes - sounds good.
I was talking to Chris V on the phone this week, he's been following your thread with interest. He suggested another way of applying the feedback, but I'll have to get him to put it down in an email (with diagram) so I can understand.

[Mike H]

Now that I've committed to putting this amp in a box, is there any way to protect a reg chip from a rogue tube?

300B arcs over in the OTL I built for Nigel....it takes out the reg chip.
13E1 arcs over in my own amp.....it takes out the reg chip.

Seems to be that a tube connected to something solid state and said tube suddenly going off on one will not turn out well for the silicon bit.

One thought that occurs to me is you could use something that is based on a high Voltage MOSFET. Which won't care if its drain is taken up to HT level.

[Ray P]

This sort of dialogue is great and would never happen over on the Transcendent forum.

As Steve has said, the trick is not to lose your way so that you can't backtrack to a known 'good configuration'.

[Nick]

One thought that occurs to me is you could use something that is based on a high Voltage MOSFET. Which won't care if its drain is taken up to HT level.

You could use a mosfet as a cascode above the CCS, but it will eat a bit of the headroom.

[Cressy Snr]

....As Steve has said, the trick is not to lose your way so that you can't backtrack to a known 'good configuration'.

Aye,
The configuration I have now is producing the best sound and the most music I have ever had from a stereo system.
Alternative feedback schemes are interesting and will no doubt be tried sometime before the breadboard becomes a 'proper' amplifier.
This single tube OTL amp, made possible by the 6C33C and the 13E1, for me delivers an astounding and entrancing performance. Those of us with 98dB and over, speakers that have flat impedance characteristics will be in sonic heaven with this amp type. I really don't know what to say, there's no describing it. It's pure music that comes out of these things when set up right. All of the hi-fi BS gets comprehensively trashed and what you are left with is the studio, the musicians and the music.

[Mike H]

One thought that occurs to me is you could use something that is based on a high Voltage MOSFET. Which won't care if its drain is taken up to HT level.

You could use a mosfet as a cascode above the CCS, but it will eat a bit of the headroom.

That's kind of what I was thinking of. Using the same reg chip underneath, and then what, a resistor divider to bias the MOSFET gate?

[Wolfgang]

Nick wrote: ↑
Wed Dec 12, 2018 2:48 pm

Mike H wrote: ↑
Wed Dec 12, 2018 7:40 am

One thought that occurs to me is you could use something that is based on a high Voltage MOSFET. Which won't care if its drain is taken up to HT level.

You could use a mosfet as a cascode above the CCS, but it will eat a bit of the headroom.

This would only protect the regulator, not the tube or the speakers in case another 200V cap goes bad at 170V B+. A zener or even better an emulation of a zener which has no negative effect on the sound is the only way to be safe in case of a faulty tube with these kind of amps.

[Nick]

Well, a DC detector and relay wound protect from a failed cap. Not sure how a zener will do any more than a cascode, but I probably don't understand how you suggest using it. I would have thought the rate of cap failure would be low as long as its rated properly. A fast blow Fuse in the B+ would add a bit more protection.

[Wolfgang]

That’s why Steve’s experience with the blown cap/arcing tube was a wake- up call for me. Add to it that these kind of amps really respond very accurately to whatever is done to them no matter where -and you don’t want to ruin the SQ! - and you find yourself very quickly stuck between a rock and a hard place.

Fuse in the B+ psu would work but would have to be fast acting and that means a very thin wire noticeable as “brighter, harsher” sound. This would protect the tube only against redplating caused by regulator failure or the pos. grid topic as discussed in the 6C33 OTL thread. This part already works with my protection circuit which has no fuse, no influence on the SQ. The monitoring circuit only draws 10mA from the B+ psu and is absolutely not noticeable in SQ.

If there is a faulty tube (or a good tube fails after many hrs of stressful use) and the regulator is only damaged but not fully destroyed (best case here would be a short to ground) the current will not necessarily rise high enough to blow the fuse/trigger a relay in the B+ psu but the cathode will be at B+ potential. Now the cap is the only thing between B+ and the speakers that keeps them from being destroyed by DC.

With a zener between regulator input and ground we would bypass the regulator (damaged to some degree, destroyed, or still functional if it can handle B+) and trigger the relay. It would put the cathode to ground and protect the caps/speakers. Even without fuse/relay in the B+ psu it will blow (hopefully) the mains transformer fuse. If the breakdown voltage of the zener has been chosen to be lower than max cap voltage it will keep the caps functional and thus protect the speakers. Even an old, faulty cap wouldn’t be a problem because the relay would trigger in less than 10ms during which the voltage couldn’t have risen high enough at the output in order to destroy the speakers. Biggest problem here is the leaking before the breakdown point which is very audible because at the cathode we also deal with AC/music signal.

The solution would be again a monitoring circuit that draws very little current at the cathode and doesn’t affect the regulator’s very high impedance for AC but would bypass the regulator at a very precisely adjustable voltage (no leaking like with a zener) which would allow to use also lower voltage rating caps as long as they are high enough above normal cathode DC level. This would instantly trigger the current monitoring circuit at the B+ level in case of tube failure (gassy tube, short between anode and cathode etc). I have already some ideas but am not sure how it will affect the regulator or the SQ. Will find out soon.

[IslandPink]

Just to throw in a random idea, is there anything you could do with a Hall effect sensor and a relay ?
I suppose the problem would be how to characterise a fault condition from some normal operating extremes.

[Wolfgang]

I cannot see at the moment how we could integrate the hall sensor for monitoring the current. But maybe there is a way? The digital output could trigger a SSR. I use CMOS Hex inverters for this in my circuit.... so it would go in the same direction.
A very simple but not very precise/practical solution which also works with a magnet/electro-magnetic field is the following: Take a reed relay and wind some lacquered copper wire around its middle part. For 500mA it would need about 100 windings to trigger the reed relay which would trigger the main power relays.Of course we wouldn't want to choke B+ so the wire would have to be at least the same gauge as the wire of the DC chokes.
In order to get this thing to work exactly where we want it to work it needs some patience for finding the right number of windings. Not to mention that it looks like dinosaur.

[Ray P]

Just for context, here's the modification to the 300B SE-OTL, with the zeneer to protect the regulator, that Rozenblit introduced. This part schematic is from the Tubes and Circuits book that contains the original 300B SE-OTL schematic, and reminded me that in the original circuit the feedback was taken from the cathode connection and it was later that he moved it to the other side of the output cap. I must try to find the posts on the TS forum tht introduced the change but, IIRC, it was related to the introduction of the 25Hz 300B filament supplies.

[Cressy Snr]

That’s why Steve’s experience with the blown cap/arcing tube was a wake- up call for me. Add to it that these kind of amps really respond very accurately to whatever is done to them no matter where -and you don’t want to ruin the SQ! - and you find yourself very quickly stuck between a rock and a hard place.

Depends on your attitude to risk. I've been completely risk averse for years, but these days I'm a bit more laid back. As long as these amps are used then switched off and not left unattended, I would think the risk from them is low. It was only because I put in a suspect tube that the reg and cap got trashed, otherwise the amp would have kept cruising along. Trouble is, you can't legislate for a tube throwing a fit; it's one of the hazards of using tube amps.

The thing that makes me most uneasy with OTLs is solid state rectifiers giving instant HT. I don't like it at all, it makes me cringe every time I throw the power switch.

The 6336A regulator tube data sheet specifies a minimum cathode heating time of 30 seconds before HT should be applied.
The 13E1 and 6C33C have massive >30W cathodes just as does the 6336A, so one addition I want to make to the breadboard, is a separate switch to bring in the HT when the heaters have reached full operating temperature. All it needs is a switch in the AC side of the rectifier bridge; a small price for peace of mind.

[Nick]

Just FYI, I have used these in the past to disconnect HT. Specifically in the ref phono if someone unplugs the connection between the two boxes CE specifies how log a voltage can remain on the pins. I detect the removal and use these to isolate the caps in the power supply while they discharge. Just letting them discharge on their own would take too long.

https://www.digikey.co.uk/product-detai ... ND/1280742