SKYLINER build - inspired by ODS #124 and #060

[talbany]

I think we are generally good to go at this end, if I understand it well though there may be a situation where the positive feedback is not audible and so I can double check this (lift the NFB and then reattach, with a signal)

Its been my experience, sending positive feedback through the 2nd input of the PI in an ODS's has almost always resulted in an audible noise. If not you would most likely notice an unstable bias supply reading (or red plate) due to high frequency oscillations and the amp would sound like crap!.. I can only remember a small handful that were inaudible..

Enjoy the new amp!!

Tony

[Stephen1966]

I think we are generally good to go at this end, if I understand it well though there may be a situation where the positive feedback is not audible and so I can double check this (lift the NFB and then reattach, with a signal)

Its been my experience, sending positive feedback through the 2nd input of the PI in an ODS's has almost always resulted in an audible noise. If not you would most likely notice an unstable bias supply reading (or red plate) due to high frequency oscillations and the amp would sound like crap!.. I can only remember a small handful that were inaudible..

Enjoy the new amp!!

Tony

Thanks Tony, no I think we are okay. I had some dodgy plate voltage readings earlier on but I think that was just the way I was trying to use the probes and them not making a good contact with the pins. It doesn't sound too bad as it is but I'm afraid of cranking it up with my Celestion speaker and now, it has a dummy load in its place. The heater is glowing a bit brighter in the first output tube (could be because it was a different batch from the others - it is) or its seeing all the current for the tubes down the line as well, but no red plating. I would expect positive feeback to "hollow out" the sound at least if it didn't scream blue murder and generally, there are some sweet sounds in there. It needs tuning up, for sure, but there isn't much point dialling it in until the big speaker is in a cab and I'm able to hear its full potential. I need to get my woodworking head on for that

[rootz]

Generally the procedure for hooking up the GNFB wire is??..Leaving it disconnected during the initial light off..After you get the amp up and running and put out any fires touch the wire to the speaker jack (Like Andy states) the volume should drop. If it's backwards and squeals? swap the plate leads (leaving enough wire to do so) and solder the wire to the jack and re-bias the amp and your good to go. This generally avoids the startle factor catching you off guard knocking something over trying to turn the amp off or damaging a speaker scaring the wife,dog kids etc etc..(It's usually the last wire I solder)

A little tip I learned from Emmit Clarks Repair shop

Tony

This! Getting the secondaries right is really that easy and also the method I use and recommend. Oh and an ODS will bark like crazy if you get it the wrong way around (at least mine did). And Tony's (and my) method is safe: you'll always first check the load is properly connected this way, before you possibly try to put a lot of power through a possibly unloaded OT.

On my latest amp I wired in a few freewheel diodes from output plates to ground for extra protection. You could also use a switched jack as main output jack. Connect the switch to ground: always a load. Or do the Hiwatt trick (think of this one as an improved version of the aforementioned solution), but then you'd need Cliff jacks for the output. Both work, but only if nothing is plugged into (one of the) the outputs! Without a solid connection, well.....

[Stephen1966]

On my latest amp I wired in a few freewheel diodes from output plates to ground for extra protection...

How would that work?

edit: polarity wise... I'm not not clear about the protection it would offer. Surge? Load - somehow?

[bepone]

do we have any sound here?

[rootz]

On my latest amp I wired in a few freewheel diodes from output plates to ground for extra protection...

How would that work?

edit: polarity wise... I'm not not clear about the protection it would offer. Surge? Load - somehow?

A better, more common name, would probably ‘flyback diodes’. They protect in situations where no current is flowing in the OT and thus voltages are sky high. The diodes are reverse biased and connected between the ot secondaries and ground. If a voltage spike occurs due to a load not being present or suddenly disconnected, the high voltage spike will bleed off through the diodes because you exceed the maximum reverse voltage. Yes, this will blow the diodes, but might just save the OT. Might, you say? Yes, because it might not work for every frequency. This depends on the characteristics of the diodes used. Ken Fisher supposedly claimed the diodes also have some effect on the sound of the amp: some high end bleeding off. Never dived deep into the matter to know for sure if both points are valid and to what extent they have implications. But with most things, there is no free lunch.

Another possibility to always provide some kind of load: a string of resistors (220k total) from the ot primaries to ground ánd a 220r resistor one the secondaries. Iirc Gingertube at diyaudio uses this solution. The exact values of the resistors can be calculated. The string on the primaries is probably needed because you easily exceed the maximum working voltage of the resistors with 2xB+ (full swing) at the anodes of the output tubes.

But then again, there is simply no substitution for using the best fail safe: your brain aka common sense. If you’re amp is suddenly silent, or there’s no sound after power on in the first place, don’t continue playing with the master on 7. That’s just stupid and asking for magic white smoke and/or unintended welding of costly copper.

[Stephen1966]

A better, more common name, would probably ‘flyback diodes’. They protect in situations where no current is flowing in the OT ...

So, forgive me, I'm still trying to get my head around this, in a practical sense this could offer some protection for the OT if for instance, the speaker jack were not plugged in? It's possible I guess. I mean, forgetting to plug it in. Has anyone ever done that? A whole new topic perhaps: "The Day I Forgot to Hook Up My Cab"

[Stephen1966]

do we have any sound here?

Yes it's producing music if that's what you can call my playing.

[Stephen1966]

...

Another possibility to always provide some kind of load: a string of resistors (220k total) from the ot primaries to ground ánd a 220r resistor one the secondaries. Iirc Gingertube at diyaudio uses this solution. The exact values of the resistors can be calculated. The string on the primaries is probably needed because you easily exceed the maximum working voltage of the resistors with 2xB+ (full swing) at the anodes of the output tubes.
...

Is this the one you are thinking of?

https://www.diyaudio.com/forums/attachm ... bias_a-jpg

[Richard1001]

On my latest amp I wired in a few freewheel diodes from output plates to ground for extra protection...

How would that work?

edit: polarity wise... I'm not not clear about the protection it would offer. Surge? Load - somehow?

A better, more common name, would probably ‘flyback diodes’. They protect in situations where no current is flowing in the OT and thus voltages are sky high. The diodes are reverse biased and connected between the ot secondaries and ground. If a voltage spike occurs due to a load not being present or suddenly disconnected, the high voltage spike will bleed off through the diodes because you exceed the maximum reverse voltage. Yes, this will blow the diodes, but might just save the OT. Might, you say? Yes, because it might not work for every frequency. This depends on the characteristics of the diodes used. Ken Fisher supposedly claimed the diodes also have some effect on the sound of the amp: some high end bleeding off. Never dived deep into the matter to know for sure if both points are valid and to what extent they have implications. But with most things, there is no free lunch.

Another possibility to always provide some kind of load: a string of resistors (220k total) from the ot primaries to ground ánd a 220r resistor one the secondaries. Iirc Gingertube at diyaudio uses this solution. The exact values of the resistors can be calculated. The string on the primaries is probably needed because you easily exceed the maximum working voltage of the resistors with 2xB+ (full swing) at the anodes of the output tubes.

But then again, there is simply no substitution for using the best fail safe: your brain aka common sense. If you’re amp is suddenly silent, or there’s no sound after power on in the first place, don’t continue playing with the master on 7. That’s just stupid and asking for magic white smoke and/or unintended welding of costly copper.

I always thought induction voltage spikes have reversed polarity and these would be shunted by the diodes that are placed in reverse polarity. Much like a diode used on a relay coil that shunts the induction voltage. I have seen diodes used on output transformer primary's that have a reverse breakdown voltage of 3KV to 15KV so if this voltage has to be exceded to offer any protection, the transformer insulation would most likely break down first.

A more easy way for 'no load' protection was done by fender. Simply conecting the speaker output to ground when no plug is connected (much the same way the input of the amp is wired) Some others dit the same thing, but instead of wiring the output straight to ground (if no plug is connected), a small lightbulb is placed between the jack switch contact and ground. Not to produce a nice light effect, but just to offer some resistance to ground instead of a short.

[bepone]

do we have any sound here?

Yes it's producing music if that's what you can call my playing.

nice! share some sound example to see what kind of beast is born!

[Stephen1966]

...
A more easy way for 'no load' protection was done by fender. Simply conecting the speaker output to ground when no plug is connected (much the same way the input of the amp is wired)...

Do you perhaps mean something like this?

ODS 124 output jacks.jpg

[rootz]

My explanation off the concept of reverse biased diodes is flawed I see (or must I say plainly wrong). I blame it on my ageing brain.

In short and now correct:
In a push pull amp one tube goes high and the other goes low voltage at the same time, right? So if you have a 100V p-p voltage swing at the plates and a B+ of 450V, plate one is at 550V and the other at 350V at max swing. Again with a load connected. If one tube drops it’s plate to 60V when saturating, the other plate must go up by the same amount, thus reaching 840V peak (390+450). And that is into a resistive load. Speakers are not entirely resistive and cutoff of the output tubes is rather abrupt, so peaks in excess of 1kV are not uncommon in overdriven power amps. Still not to big of a deal with winding insulation that is made to withstand that voltage.
Now imagine a load suddenly being disconnected. The OT is essentially a very large inductor and thus stores energy. But because there is no load, the load becomes an infinite resistance and therefor the reflected primary impedance becomes infinite. Because the transformer keeps the current going, the voltage must go up at one plate. And because it is a push pull stage, the voltage at the opposing plate needs to drop below 0V. Here the diodes come in, as the start to conduct as the plate cannot go further below 0V than the (combined) forward voltage drop of the diodes. This limits the voltage spike on the other plate.
I can see how this would influence the sound of a overdriven power amp, damping out peaks and spikes. Supposedly it makes the power amp a bit smoother. Unless you’re Joe Bonamassa you probably run you power amp clean on your ODS. Therefor I don’t see much of a sound penalty for much added safety. Keep in mind that Ken Fisher was not a fan of protection diodes like these, but his amps were typically not running the power amp clean.

IMHO the diodes provide better or at least substantial added protection than switching output jacks like Fender (and Dumble, at least in the layouts floating around here at the forum) used. Those only work if nothing is plugged in the output. If you have a bad connection or cable, or the cable suddenly becomes disconnected, you are out of luck. Especially when you strike some good power chords at full output volume. Fenders protection is what it is: a protection against no load.

The schematic posted inbouw reaction to my first post shows Baby Huey. That amp has no protection diodes or resistors. The resistors from the plates of the output tubes are there as part of a so-called Schade feedback network. A better term would be shunt feedback. It is nothing more than an anode follower output stage where the output impedance of the preceding stage forms a feedback network with the resistors from the output plates.

[Stephen1966]

Thank you, for the update rootz. A good lesson for me here but alas, you are not the only one with an ageing brain. I kinda follow these kind of discussions but I'm particularly slow on my ability to conceptualise the ideas, given that a lot of the ideas are still brand new. I really need schematics, logical paths illustrated, or layouts from which I can derive these logical pathways. I used to have a very strong sense for the mathematical when I was younger, but these days, it's my visual sense and imagination which is stronger. Which is to say, if I can see, I can understand it - at least more easily. There's no accounting for any actual lack of ability, unfortunately.

I can visualise reversed bias diodes to ground and the contexts you describe both in fault protection and normal operation. Again, a schematic or layout of some description would help but it's going to be necessary to do a fair bit of background reading to fully grasp both what you are describing, and coming up with an implementation that can be fitted in this, or any other similar amp. That said, you will see I have used flyback resistors in my relays so I have a "rudimentary understanding" of what they're doing. You've certainly done your bit with the introduction to the theory though, I thank you for that. And please, don't be deterred, if you have any similar contributions to make on anything else, please keep it coming. It means a lot to me that you guys are so patient with such a dullard as I.

I'm happy with the Fenderish no load protection in the execution of the output stage in the 124 as described, and emulated in my design as well. A little piece of mind there, because of all these things likely to go wrong, forgetting to hook up the cab is probably top of the list.

Affirmative: the schematic I found was of the Baby Huey. But "Schade feedback network/shunt feedback" = another addition to reading list there, for another day perhaps. I'm stumbling around in the dark on Gingertubes' resistor based protection circuit though, perhaps some pointers on where to find it? A string of resistors (contrasted with diodes that might blow out) may at least allow the show to go on.

[Stephen1966]

do we have any sound here?

Yes it's producing music if that's what you can call my playing.

nice! share some sound example to see what kind of beast is born!

Jo! I actually have a plan there... I don't have any recording equipment here at home (except for a mobile phone - foy!) but I have a friend with his own studio and I'm doing some lutherie work for him, in return for studio time. He's a true friend, we would do these things for nothing in return, but we're both that fair-minded. In time then, lots of stuff to do before we can be ready for that, but I have to ask, are you sure? The expression would be: "Careful what you wish for!"