Protecting our Amps

After reading the thread A bright flash, a low dull thunk and diminished volume
I got thinking about ways to protect our amplifiers.

From my experience with other tube gear the concern tends to be damage to
the filaments and stressing the transformer when you first turn the device on.
One way of protecting the gear is by the use of a "Slow Start".

A Slow start is basically a resistor that is put in series with the input of the
power transformer with a time delay relay that will bypass the resistor after
a few seconds then applying full line voltage to the power transformer.

The Slow Start lets the tubes start to warm up before hitting them with full
voltage that in turn puts less stress on them and on the power transformer.

As I understand it a Thermistor does the same thing but with with a voltage drop.
So it may be worthwhile to install a Thermistor to help protect from the initial inrush.
Note: It would seem that a Thermistor would help extend the life of all the tubes.

-

Another concern tends to be when the AC Power cycles off and on again.
Using something like This (GFCI Plug with Surge Protection)should
help since it would require a manual reset after the power outage.
(That info came from a linked post on the diode mod thread)

-

Lastly is the Rectifier tube, I would tend to think if you used Thermistor
that you could then switch over to SS rectification with no ill effects.
(Please correct me if I'm wrong here).

But if you still want to use Tube rectification you could
get a higher rated tube but they can be pricey.

Another option would be to do the Tube rectifier diode mod but a post in that thread
suggested tying pins 4 and 6 together to double the forward current capacity of the 5AR4
actually using two solid state diodes per leg of the PT HV secondary before the 5AR4 and
also to tie a .05uF 600V or 1000V cap to ground from the joined #4 & 6 pins.
(Read the thread and the linked threads as they contain some good info).

-

It would seem that for a few dollars you should be able to protect your amp

Hopefully this info is useful to others.

A thermistor is a great idea. I think a lot of people are using them when using SS rectification as you mentioned. It helps deal with the B+ overvoltage before the tubes start conducting. Same concern as in that other thread. I think it also makes the filaments heat up more slowly though, which might go against this benefit since we are trying to get the tubes conducting. Not sure, but I hear overall they are effective.

I'm a fan of the diode mod too, and have used it on a few amps where I still want to keep a tube rectifier. I think it helps with the reverse voltage protection and could potentially extend the life of the tube, but even if it doesn't it is so cheap it seems worth doing just in case. With the mod in place you can also go one step further and parallel the plates so the tube essentially becomes a single diode with less voltage drop (Dave Gillespie mod). It can help a weak tube supply a power hungry circuit. Basically all the tube is doing at that point is delaying the B+ onset and passing the electrons through a hot vacuum, which of course makes them happy

Personally I prefer SS rectification for HiFi in terms of sound quality, but there is something nice about a big toasty rectifier that is hard to let go.

The off/on issue (I have blown many a rectifier tube) I solved by designing and building a power protection/power sequencer. First, on initial power on, every outlet comes on at the same time. Since tubes take time to warm up anyway, there is little risk of a power problem here. I stay away from SS rectifiers, I just like the tubes better. Now comes the protection. On power off, only the power amp (my st-120) turns off first. Then, about 25 seconds later the pre-amp shuts off, and then another 10 seconds or so and any other equipment powers off. This off sequence protects first, the speakers for a pop when source pre-amp powers off, and pre-amp from transients when other sources power off.

Lastly and most important of all, if the power switch is turned on immediatly after turning off, nothing happens. One of the lights on the power controller flashes rapidly indicating you did the dumb thing. Power cannot be cycled back on for about 4 minutes. This allows caps to discharge enough and filaments to cool enough that a power on is about normal.

My current controller uses some relay timer modules but, I am designing a better one using an arduino with possibly some more features like brownout cutoff and power company glitch protection to absolutely prevent the off/on scenario. I have found a cheap power box that has a ton of space inside to build pro looking controllers. Even cheaper getting used ones on ebay (sorry, not giving it away until I have scarfed up enough to build units for my friends). Designs will be available when I finish. I am building two different models.

Yes, yes, this problem is known to anyone who has to do with tube amplifiers, including us from AudioValve.

For this reason, we have developed a printed circuit board that supports this softstart for the power transformer. In principle it is as you have already described.

An NTC limits the inrush current and if it is then hot and low impedance, the heating voltage is completely built up on the secondary side and then used to short-circuit the NTC with the help of a relay.

So it is a feedback loop fed from the secondary side of your own power transformer - in our case, the rectification for the precursor heating (12.6 VDC) and that works great and costs no effort.

We have soldered this circuit board right on the back of the power socket, since this is the best placed there.

Before I forget it, we also made a phase finder on it, so when the green LED on the circuit board lights up, you know that the phase at the mains socket is correct.

regards, Helmut

I too have struggled with the issue of how to delay the shock of rapid current flow in tube amps. The thermistor does work however it is extremely important to make sure it is located away from any wiring or plastic. The do get very warm and can melt wiring easily. They do offer the advantage of dropping supply voltage a little, but again at the expense of heat. One of the undiscussed issues with diodes verses tube rectifiers is the higher B+ voltage. Many of the amps I restore were built when line voltage was 110-115 AC. Typically in my shop it ranges from 121-124, with some drops to 120 when everyone in town has air conditioning on. I won't go through the math for you but if your amp uses 450 volt caps and operated at about 400 after warm up, with this increased line voltage you are stressing them. This is often the case with Dynaco amps where they may operate within 25 volts of rated value. Sometimes you can use different rectifier tubes to lower B+ (like a 5R4 instead of a 5AR4) or you can use a bucking transformer, if you have room for mount it. The advantage of the bucking transformer is that it also drops the filament voltage which saves tubes. Best thing is to use a combination of things and use a circuit to ramp up voltage slowly either the AC or the DC. There are a number of aftermarket power supply power conditioners that bring up AC slowly. The can be expensive, but they will protect your amps and equipment.

Bruce

Bob Latino has posted that a thermistor on the power transformer primary would do little to protect a 5AR4. I wanted a second opinion, so I asked knowledgeable tube distributor Jim McShane the same thing. He agreed with Bob that it would do "little to nothing" (with regard to the 5AR4).

However, I also asked Jim about Eli Duttman's suggestion of putting a thermistor in series between pin #8 of the 5AR4 and the first power supply capacitor, in addition to doing the diode mod. Jim agreed that adding a thermistor after the 5AR4 would help protect the tube rectifier.

Given a CL-90 is about $2.50, that seems like very cheap insurance, compared to a $20 to $40 5AR4. I'll consider adding them to my ST-70s and MK-IIIs the next time I have them on the bench.

it is really so that you can install this NTC at different points in the circuit on the secondary side too, you just have to be aware of it - when high current flows - a series resistor is installed, which then becomes hot, if he is not bridged over a relay after a certain time. In any case, I also welcome a limitation of switch-on current.
thx, Helmut

.

I like the idea of bridging the thermistor with a relay after the initial protection period. This ensures that it is ready to do its job again if the power is cycled as well as reducing dissipation to zero in normal operation.

1973shovel wrote:

However, I also asked Jim about Eli Duttman's suggestion of putting a thermistor in series between pin #8 of the 5AR4 and the first power supply capacitor, in addition to doing the diode mod. Jim agreed that adding a thermistor after the 5AR4 would help protect the tube rectifier.

Would a thermistor work in this application? There is relatively low current in the B+ circuit compared to AC primary. I am wondering if they would heat up enough to change resistance properly. CL90 for example requires at least 0.5 amp. Maybe there is a special low current model?

rjpjnk wrote:

1973shovel wrote:

However, I also asked Jim about Eli Duttman's suggestion of putting a thermistor in series between pin #8 of the 5AR4 and the first power supply capacitor, in addition to doing the diode mod. Jim agreed that adding a thermistor after the 5AR4 would help protect the tube rectifier.

Would a thermistor work in this application? There is relatively low current in the B+ circuit compared to AC primary. I am wondering if they would heat up enough to change resistance properly. CL90 for example requires at least 0.5 amp. Maybe there is a special low current model?

If you're referring to the CL-90's data sheet showing only 25% (0.5A) 50% (1A) 75% (1.5A) and 100% (2A) these are to indicate ranges of resistance at current percentages. The 0.5 amp shown isn't a minimum at which the thermistor will function.

As an example, I have an SCA-35 to which I've added a CL-90 directly after the rectifier diodes. The total B+ current draw of all the tubes is approximately 240 mA (0.24A) maximum, per the tube data sheets. The CL-90 functions under that current draw, and slows the B+ surge. There is a calculated resistance on the CL-90 of about 16 ohms remaining at that current after warm-up, which drops my B+ voltages to match Dynaco's readings nearly exactly.

PeterCapo wrote:Would I be correct in assuming they meant a thermistor in the PT primary would not help protect the 5AR4 in terms of failure from forward current demand during [otherwise] normal operation?  Putting a thermistor at the output of the 5AR4 would limit the forward current some - you'd still get some voltage drop after it heats up.  But, Dave Gillespie’s mod that ties the plates of the 5AR4 together could also be a significant improvement.  I posted about this mod in a sticky in this forum - start at Post n°64 here https://dynacotubeaudio.forumotion.com/t1006p50-tube-rectifier-diode-mod

Adding a thermistor after the 5AR4 in a Stereo 70 or Mark III is different in the sense that neither amplifier puts bone-crushing stress on the 5AR4 to begin with.  The Mark III actually stresses the 5AR4 less than the ST-70.  You could add a thermistor after the 5AR4 just the same, but then you'd want to see what happens to the voltage levels on the B+ nodes downstream.  IMO, I'd want them to still line up with the original design values.

Your assumption regarding a thermistor on the primary is the same as mine, in as far as protecting the 5AR4 goes. Thank you for the link to yet another "Gillespie Mod". I had never seen that recommendation, and will have to give it a more in-depth look.

Regarding the 5AR4 in a ST-70, to quote Gillespie, "A much bigger compromise -- in my opinion -- is the use of a single rectifier tube to support a 70 watt amplifier. Competitors with competing products from the likes of Altec, Fisher, and Pilot all used dual rectifier tubes in their products of similar power levels. As a result, there is no doubt that the rectifier tube in an ST-70 is the hardest working tube in the chassis, and often will show the highest failure rate as well -- even when using NOS examples."

Which is why I was considering adding a CL-90 after the 5AR4. It also has the added advantage of dropping the voltage by a few volts, which is almost never a bad thing, given today's generally higher AC at the wall, which is why I was thinking about it for my MK-IIIs (although a variac or bucking transformer would probably be better). But with the 5AR4 modification you've linked to, the thermistor after the 5AR4 might become a moot point.

.

PeterCapo wrote:I need to refresh my memory of the landmark Gillespie article.  But, didn't he also indicate that the limitations of the single 5AR4 in the Stereo 70 really only show up when the amplifier is running full tilt?  Meaning, is it really an issue if the amp is running at a more moderate level?  My impression is that some of the less robust current production 5AR4 may not hold up even at more moderate levels, but I would not think it should be a problem with the more robust current production Tung Sol and Gold Lion 5AR4/GZ34..?

I would have to refresh my memory too, but I believe you're correct. Rather than going to elaborate measures, realistically I'll likely do the "yellow sheet diode mod", and call it good. I've actually only had one 5AR4 fail on me in a ST-70, and this was prior to adding Roy's VTA input stage. A brand new Phillips 5AR4 arced on me upon first insertion, and this was back in the 1990s. I've never had any issue with the Sovteks in my ST-70, even without the diode mod. I tend to be a belt and suspenders kind of guy, which is why I was considering a thermistor after the rectifier.

.

PeterCapo wrote:Not to belabor the point, you are aware that Yellow Sheet Diode Mod does nothing to protect the 5AR4 from forward current demands?

I don't think you're belaboring the point at all, and hopefully this information will be valuable to everyone. To answer your question, yes, I'm aware that the diode mod won't help regarding current demands, only to relieve the PIV on the 5AR4. If I were experiencing 5AR4 failures then I would have to take further measures, but since I'm not, the $1 for two UF4007s seems prudent. The thermistor  I was considering indicates my previously mentioned "belt and suspenders" approach to things, and is probably unnecessary for my application. "If it ain't broke..." and all that.

Thank you again for that interesting link. It's always a pleasure to read your posts.

1973shovel wrote:

rjpjnk wrote:

1973shovel wrote:

However, I also asked Jim about Eli Duttman's suggestion of putting a thermistor in series between pin #8 of the 5AR4 and the first power supply capacitor, in addition to doing the diode mod. Jim agreed that adding a thermistor after the 5AR4 would help protect the tube rectifier.

Would a thermistor work in this application? There is relatively low current in the B+ circuit compared to AC primary. I am wondering if they would heat up enough to change resistance properly. CL90 for example requires at least 0.5 amp. Maybe there is a special low current model?

If you're referring to the CL-90's data sheet showing only 25% (0.5A) 50% (1A) 75% (1.5A) and 100% (2A) these are to indicate ranges of resistance at current percentages. The 0.5 amp shown isn't a minimum at which the thermistor will function.

I was referring to the specified minimum current as shown in column 11 of this datasheet.
https://www.mouser.com/datasheet/2/18/AAS-920-325D-Thermometrics-NTC-Inrush-031814-web-1315885.pdf

It says 0.5A is the min, but I am glad to hear it is working fine at lower currents and only adding 16 ohms resistance.

rjpjnk wrote:I was referring to the specified minimum current as shown in column 11 of this datasheet.
https://www.mouser.com/datasheet/2/18/AAS-920-325D-Thermometrics-NTC-Inrush-031814-web-1315885.pdf

It says 0.5A is the min, but I am glad to hear it is working fine at lower currents and only adding 16 ohms resistance.

Ahh, I see what you mean. If you look at column 11 for the CL-90, it indicates ≤ (less than, or equal to) 0.5 A, so apparently anything less than one-half amp will allow the thermistor to function at its 25% (7.8 ohms) rating. When I said mine was functioning at 16 ohms, I was estimating (7.8 ohms x 2) based on my estimated current draw of 240 mA, assuming I'd get less resistance drop as current went down. Apparently they don't work that way, so thank you for making me aware of that. I never actually measured the voltage drop and calculated the actual resistance, but I will.

The good news is, they will function, or that has been my experience with my SCA-35.

1973shovel wrote:

Ahh, I see what you mean. If you look at column 11 for the CL-90, it indicates ≤ (less than, or equal to) 0.5 A, so apparently anything less than one-half amp will allow the thermistor to function at its 25% (7.8 ohms) rating.

It actually says 0.5 ≤ 1 ≤ 2, which is really confusing having what appears to be a 1 (one) in between the min (0.5), and the max (2).

I'm not sure how to interpret this, but I assume the "1" is a typo and supposed to be an "I" (for current) giving 0.5 ≤ I ≤ 2, meaning I must be greater than or equal to 0.5A, and less than or equal to 2A.

But all of this is moot because apparently they work fine based on your testing. It will be interesting to hear the measured voltage drop you measure.

rjpjnk wrote:I'm not sure how to interpret this, but I assume the "1" is a typo and supposed to be an "I" (for current) giving 0.5 ≤ I ≤ 2, meaning I must be greater than or equal to 0.5A, and less than or equal to 2A.

But all of this is moot because apparently they work fine based on your testing. It will be interesting to hear the measured voltage drop you measure.

I can say with a fair amount of certainty that the "1" on the data sheet you linked to is not a typo. I say that because I have a GE (General Electric) NTC thermistor data sheet, and it says exactly the same thing! I'm not sure how to interpret that either.

I put my meter on the CL-90 this morning, and I get a 2 VDC drop after the amp is warmed up. 2V / 0.24A (approximated maximum current draw of all tubes) = 8.3 ohms, and my data sheet for the CL-90 indicates 7.8 ohms at 25% of full load. Keep in mind that the 240 mA tube draw are maximums taken by adding up the figures on the various data sheets, meaning I'm not measuring the actual current draw. Quiescent current would be less, but apparently the thermistor will still function at that lower current, since 25% would be the 0.5A you originally stated.

Interesting exchange, thanks!

1973shovel wrote:

I put my meter on the CL-90 this morning, and I get a 2 VDC drop after the amp is warmed up. 2V / 0.24A (approximated maximum current draw of all tubes) = 8.3 ohms

2V is fantastic! Low enough for sure. Thanks for this important data.