These Diodes.???

[trem]

We can sit around and come up with "better" ways all day, but in the real world, out of the hundreds of thousands of amps out there, how many times have you encountered these cathode diodes causing a problem? Or making some other problem worse? I haven't seen it.



That may be impressive, but wiring the cathode circuit to pin 7 of the EL34s is maybe less than impressive. Not to mention wiring their heaters to pins 2 and 6.

Yeah, I have seen a similar F-Up on the AC15. The numbers were wrong, but the pictorial was correct.
You have pretty good eyes to see that ALL the way from Michigan......
Best

[Enzo]

It's a zen thing, I AM the amp.

[jpfamps]

Yeah, it's always a sifting and weighing process. Fixing one exposure opens the door to others. The trick is in finding protection schemes where the protection scheme doesn't add other failure modes.

If you're interested, I'll do a schemo up for the shutdown circuit and post it.

I would certainly be interested in any circuit noodlings for a electronics shut down on over current.

I've not used the photovoltaic optoisolators.

I did try to come up an electronic shut down circuit which would latch on in the event of over current, however it used an op amp, and thus needed a low voltage supply to work and so never got beyond the "back of an envelope" stage.

Regardless, the use of the diodes to protect the current sense resistors leaves the amp no more vulnerable to a valve short than any amp that simply grounds the power valve cathodes.

[diagrammatiks]

We can sit around and come up with "better" ways all day, but in the real world, out of the hundreds of thousands of amps out there, how many times have you encountered these cathode diodes causing a problem? Or making some other problem worse? I haven't seen it.



That may be impressive, but wiring the cathode circuit to pin 7 of the EL34s is maybe less than impressive. Not to mention wiring their heaters to pins 2 and 6.

It's not so much the diodes themselves as the wattage rating of the resistors. As R.G. said earlier, the diodes are there so they can use smaller wattage resistors.

[R.G.]

I would certainly be interested in any circuit noodlings for a electronics shut down on over current.

Here it is:
Return of the More MOSFET Follies

I did try to come up an electronic shut down circuit which would latch on in the event of over current, however it used an op amp, and thus needed a low voltage supply to work and so never got beyond the "back of an envelope" stage.

I did the same, but noted that the current needed could be made into a few milliamperes, and derived it with a zener from B+. I used a low power dual comparator and some CMOS logic. Worked as expected, where "expected" included messing with the time constant to ignore short duration trips. That was about 10 years ago. Today, I'd use a low power microcontroller to get even lower power. Some of these have internal comparators, but using the internal A->D converters and doing the decisions on "tripped" in software makes even more sense.

Once you can sense tube currents, and do software on it, the applications open up pretty quickly. The obvious ones, shutting things down on too long a current-time integral are pretty easy, but most of the uCs with A-Ds in them have many channels, so you can also run a big resistive divider and monitor B+ as well, and probably screen voltage. The decision about what to do in terms of shutting things down eventually devolves down to what to tell the guy who's trying to make noise with the amp.

You could:
- turn off the AC power to the amp
- for a four-tube amp, turn off one pair of outputs to let the show go on
- sense temps at various places
- turn on a "disaster recovery" fan
- light up LEDs discretely on the back of the amp to tell the guy what failed
- lots more

It's about here that the "Tubes are Holy" guys start jumping off the train.

[diagrammatiks]

Here it is:
Return of the More MOSFET Follies


I did the same, but noted that the current needed could be made into a few milliamperes, and derived it with a zener from B+. I used a low power dual comparator and some CMOS logic. Worked as expected, where "expected" included messing with the time constant to ignore short duration trips. That was about 10 years ago. Today, I'd use a low power microcontroller to get even lower power. Some of these have internal comparators, but using the internal A->D converters and doing the decisions on "tripped" in software makes even more sense.

Once you can sense tube currents, and do software on it, the applications open up pretty quickly. The obvious ones, shutting things down on too long a current-time integral are pretty easy, but most of the uCs with A-Ds in them have many channels, so you can also run a big resistive divider and monitor B+ as well, and probably screen voltage. The decision about what to do in terms of shutting things down eventually devolves down to what to tell the guy who's trying to make noise with the amp.

You could:
- turn off the AC power to the amp
- for a four-tube amp, turn off one pair of outputs to let the show go on
- sense temps at various places
- turn on a "disaster recovery" fan
- light up LEDs discretely on the back of the amp to tell the guy what failed
- lots more

It's about here that the "Tubes are Holy" guys start jumping off the train.

If you were a real hero you'd be able to do all that with philbricks RG.

[R.G.]

Point~! (slaps mat)

I *did* work with Philbricks at one time in my undergrad years. Does that count?

[trem]

Point~! (slaps mat)

I *did* work with Philbricks at one time in my undergrad years. Does that count?

If you do not mind me asking....what kind education/degree(s) do you have.?
Not trying to have any kind of leverage against anybody or anything like that. I am a guy that barely made it out of High school, and then have had "blue collar" jobs all my life.
All I know is what I glean from men like yourself on Forums like this, and from reading (and re-reading) books by guys like Merlin. It is just kind of interesting to me, to know what you guys have accomplished. Not everybody likes to talk about themselves however, I will not be offended if you pass on this......
Best

[R.G.]

No big secret. BSEE, MS with specialization in EE (that's what the school calls an MSEE). However I now work for a guy who is self taught in electronics, and did a bang-up job of it. Degrees are one capsule way to package a bunch of stuff you'll (likely) need. They're not the only way to get it. I know multidegreed EEs that I wouldn't trust to change batteries in a flashlight.

I should mention that the Philbricks were antiques when I messed with them decades ago.

[trem]

10-4.
Thank You and Well Done

[jpfamps]

Here it is:
Return of the More MOSFET Follies


I did the same, but noted that the current needed could be made into a few milliamperes, and derived it with a zener from B+. I used a low power dual comparator and some CMOS logic. Worked as expected, where "expected" included messing with the time constant to ignore short duration trips. That was about 10 years ago. Today, I'd use a low power microcontroller to get even lower power. Some of these have internal comparators, but using the internal A->D converters and doing the decisions on "tripped" in software makes even more sense.

Once you can sense tube currents, and do software on it, the applications open up pretty quickly. The obvious ones, shutting things down on too long a current-time integral are pretty easy, but most of the uCs with A-Ds in them have many channels, so you can also run a big resistive divider and monitor B+ as well, and probably screen voltage. The decision about what to do in terms of shutting things down eventually devolves down to what to tell the guy who's trying to make noise with the amp.

You could:
- turn off the AC power to the amp
- for a four-tube amp, turn off one pair of outputs to let the show go on
- sense temps at various places
- turn on a "disaster recovery" fan
- light up LEDs discretely on the back of the amp to tell the guy what failed
- lots more

It's about here that the "Tubes are Holy" guys start jumping off the train.

Ah, I missed the "Return of the MOSFET Follies".

I was not 100% happy about about running opamps of a stepped down B+ as I though it slightly wasteful of power.

Thinking about this last night, it might be better to power the over current detection/ shut down circuitry from the bias supply (in a fixed bias amp), and this could be done in such a way that if the bias supply fails the amp also switches the HT off.

Tubesync's Bias Engine use a PIC controller to monitor bias currents etc, however this seems to lift the cathode connection (TubeSync - Guitar Tube Amp Biasing, EL34 s valves / tubes and more) rather than kill the HT.

[R.G.]

Ah, I missed the "Return of the MOSFET Follies".

Actually, you didn't. I put it up just before posting the note to you.

I was not 100% happy about about running opamps of a stepped down B+ as I though it slightly wasteful of power.

Yep. There are a number of details to be worked out to make a real design solid.

Thinking about this last night, it might be better to power the over current detection/ shut down circuitry from the bias supply (in a fixed bias amp), and this could be done in such a way that if the bias supply fails the amp also switches the HT off.

... and that's one of them. I think the right thing to do might be to run the protection setup on a tiny parasitic power supply from the AC mains, kind of an internal wall-wart, so it's always alive when the power plug is in the wall. The power can be made vanishingly small.

Tubesync's Bias Engine use a PIC controller to monitor bias currents etc, however this seems to lift the cathode connection (TubeSync - Guitar Tube Amp Biasing, EL34 s valves / tubes and more) rather than kill the HT.

I don't know what they have inside there, but that seems reasonable. I designed up a microcontroller auto-biasing system for one of the Workhorse amps, but we decided to not use it when the red-green light bias system turned out so easy to use. On this system, I used a floating MOSFET in the plate supply for shutdown, but it was driven from an ordinary optoisolator and floating circuitry on the MOSFET gate drive. That's a variation I didn't put in "Return..." but it's easy enough to do.

I opened the cathode connection for my original overcurrent detector/protector too. I suspect that if the Tubesync guys or any of their customers read this forum, we might see isolated LED/PV MOSFETs in the plate supply from them next.

Once you have made the step to having a controller in there, the number of add-on applications explodes. Lose bias? Turn off the B+ and light up the "lost bias" LED. Or put "bias missing" in the diagnostic display. Tube overcurrent? Note which one did this in the display. Or spit out the conditions noted in RS232 form on diagnostics plug, like all cars do today.

There are lots of places in a tube amp to monitor for the right voltage being there, and few enough to monitor with a small micro. For instance, $2 PICs can have up to 12 A-D inputs. One could use a high-impedance resistive divider on the plates of all the tubes, detect absence of signal to figure out when to read the voltages, and figure out which tube(s) were having problems.

It's a target-rich environment.

[jpfamps]

Actually, you didn't. I put it up just before posting the note to you.

Yep. There are a number of details to be worked out to make a real design solid.

I was looking into the current limiting on the HT (unless you employed foldback current limiting) and came to the conclusion that it probably wasn't a great idea as a means of protecting the mains transformer.

The reason I came to this conclusion was that in a fault condition, which I think is most likely to be a shorted valve, then the current limit employed would still be significant and stress the series pass element (making it unlikely that the series pass element could survive the short for any appreciable length of time), but would also make the clearing time of the HT fuse longer. That is of course assuming that a quick blow fuse was employed in the HT after the main filter caps, rather than a slow blow fuse on the HT winding before the rectifier. Shutting down the HT seemed the better option.

However, a current limit on the screen supply could well be worthwhile as the current could be limited would be much lower, and the series pass element would be able to survive.

I think the right thing to do might be to run the protection setup on a tiny parasitic power supply from the AC mains, kind of an internal wall-wart, so it's always alive when the power plug is in the wall. The power can be made vanishingly small.

A small transformer maybe, or even a capacitor coupled supply (CE might not like that....).

I don't know what they have inside there, but that seems reasonable. I designed up a microcontroller auto-biasing system for one of the Workhorse amps, but we decided to not use it when the red-green light bias system turned out so easy to use. On this system, I used a floating MOSFET in the plate supply for shutdown, but it was driven from an ordinary optoisolator and floating circuitry on the MOSFET gate drive. That's a variation I didn't put in "Return..." but it's easy enough to do.

I opened the cathode connection for my original overcurrent detector/protector too. I suspect that if the Tubesync guys or any of their customers read this forum, we might see isolated LED/PV MOSFETs in the plate supply from them next.

Once you have made the step to having a controller in there, the number of add-on applications explodes. Lose bias? Turn off the B+ and light up the "lost bias" LED. Or put "bias missing" in the diagnostic display. Tube overcurrent? Note which one did this in the display. Or spit out the conditions noted in RS232 form on diagnostics plug, like all cars do today.

There are lots of places in a tube amp to monitor for the right voltage being there, and few enough to monitor with a small micro. For instance, $2 PICs can have up to 12 A-D inputs. One could use a high-impedance resistive divider on the plates of all the tubes, detect absence of signal to figure out when to read the voltages, and figure out which tube(s) were having problems.

It's a target-rich environment.

I know very little about PIC (in fact nothing), but it seems it may be the way to go.

If you were looking for a HT interupt "module" that could be retro fitted you need to have two HT lines for amps that use differing screen and anode supplies, eg the SVT-style amps, especially as it is this type of amp that would benefit most from the extra protection.

[R.G.]

I was looking into the current limiting on the HT (unless you employed foldback current limiting) and came to the conclusion that it probably wasn't a great idea as a means of protecting the mains transformer.

That's true. And it's only one of the issues involved in coming up with a comprehensive protection scheme.

In doing any protection setup, you have to be really clear what you're protecting, and what you're willing to sacrifice (if anything) to protect parts in priority order. It's great if one solution protects everything, but you may not always be so lucky.

Case in point: the current limiter. It's not for protecting the power transformer. It's for softening the sudden current pulse on coming out of standby on the rectifier tube, and secondarily for limiting the inrush current stress on the first filter cap. In the AC30 particularly, the current inrush coming off standby is big enough to damage the rectifier tubes, it seems.

But that does complicate the issue of shutting down the amp with a shorted valve. You need to set the limit big enough to trip the AC fuse on a sustained shorted tube and small enough to protect the rectifier tube if that window is big enough. If it's not, then you need to do something to time-limit the current limit or to do a foldback. A fuse is kind of the ultimate foldback, but has this problem that it's trip point is not too predictable and varies with the size of the current overload.

A comprehensive protection setup would have to take care of the PT, OT, perhaps output tubes, rectifier tube, overheating, etc. and not have the protection schemes make a fault worse.

One issue that would definitely come up is that neither guitarists nor old-school techs are likely to appreciate some protection setup turning their amp off. The same reasoning that leads to putting aluminum foil over fuses would see the protection circuit as the problem, not the way to keep the amp from emitting black smoke. It is likely that if some hypothetical protection setup really did work and kept the amp from doing more damage to itself from loss of bias, shorted tube, shorted rectifier, etc. that the *protection circuit* would be blamed for being faulty and not letting the amp work.

EDIT: In thinking about this, it might make sense to have a "don't do nothing!" switch to tell the protection circuit to just sit and watch it burn. That would make for a great - if expensive! - teaching tool.

The reason I came to this conclusion was that in a fault condition, which I think is most likely to be a shorted valve,

I would have guessed that loss of bias was more common, but that is just a guess. That condition has a lower fault current than a shorted tube.

then the current limit employed would still be significant and stress the series pass element (making it unlikely that the series pass element could survive the short for any appreciable length of time),

True if the series pass element stays linear. A heat sink puts off the inevitable for a while till you can do something about it.

However, a current limit on the screen supply could well be worthwhile as the current could be limited would be much lower, and the series pass element would be able to survive.

That's another possibility. The same MOSFET current clamp works for screens, of course.

I know very little about PIC (in fact nothing), but it seems it may be the way to go.

PICs and their cousins-in-logic microcontrollers can be thought of as "do anything I want to, based on whatever I can sense." As long as what you want to do is reasonable to describe and you're not looking for complex actions in microseconds, they can pretty much do it.

If you were looking for a HT interupt "module" that could be retro fitted you need to have two HT lines for amps that use differing screen and anode supplies, eg the SVT-style amps, especially as it is this type of amp that would benefit most from the extra protection.

At this point I don't have amp designs cooking - too much else on my plate. But I can't stop myself from designing. It's a personal curse.

[trobbins]

Another 1R 1/4W in series with the existing 1R would act as a fuse, and another 10k from cathode to 0V would act as heater-cathode voltage limiter. But that is for diyers, as imho the designer response shows the lack of warranty repairs (for all related failure modes as he discussed) has been the justification for the choice (although no idea of how many units history he is pointing to here).