SS, 5U4, 5V4, 5AR4, 5V3 & 5Y3 Rectifiers Compared In 75 Watt Vacuum Tube Audio Amplifier 

Jennifer - I really appreciate your comment because it is my goal to put the proven and practical on RU-vid. I love theory but applied science is always my goal here. Yes, during the 60's and 70's we learned about some of these rectifiers by plugging them in and seeing what happens. We learned a lot of practical knowledge from that. With the Internet nowadays, researchers can get so many versions of the same thing and they simply all can't be right and I think it leads to a lot of misunderstanding. I wish I could document every detail and when I make errors I am called out on it by my viewers but too much detail is so laborious and even dangerous around HV and lastly, maybe just too much data. This video came to me because it is "practical" and was relatively easy to make. Since last night, I have swapped out the 7H, 200 mA choke for a 1.5H, 500 mA choke (one of the ones I showed at the end of the video) and now the amplifier will provide over 87 watts with distortion under 0.5%. Of course, I realize I am just raising the loaded voltage and lowering the Z of the PS so it just gets more powerful. I may actually be reaching the limit on resting current of the KT88's. The R of the choke that was in the amp is 126 ohms and rated at 200 mA. The bigger choke has an R of 27 ohms and it rated at 500 mA. I love this stuff and am excited to start on the 816 MV amp very soon. Lastly, you mention the Williamson design and I was not sure if the 6SN7 Williamson drivers could deliver enough undistorted peak voltage to drive a pair of KT88's to their max level but I apparently proved it is so in this video. I hope to see your nice Heathkit, Altec and JBL work posted on RU-vid soon. Stay safe and thank you for your comments.

@@powertothebauer296 ? Fender Champ will sag hard. Those 5Y3 can’t handle the in rush of power when you try to play a Les Paul into it. Massive voltage sag on that cathode bias amp.

Your description of the tube sounds like a military version of a 5R4. The 5R4 being a higher voltage, lower current rectifier, often used in ham radio transmitters with an output of 800 volts or so, would indeed, sag and clip the output more than a 5U4 as the 5R4 is a lower current rectifier. The music amplifier will typically operate at a lower voltage and higher current than the radio transmitter. You are touching on a subject that has been on my mind since uploading this video and all the comments posted. I am beginning to see that the 5U4 class of rectifiers, et al, give the guitar musician that magical sag/clipping that is so sought after and replacing the vacuum rectifier with SS may significantly eliminate that sag and clipping, basically destroying the sound that the musician wants. As for HiFi, it is the other way around. I want Bolder Dam, low impedance, high current capabilities to accurately reproduce what was originally created by the musician, sag, clipping and all in the original recording. As for damaging the rectifier filament winding on the power transformer, I agree with you, if it is a 5V, 2A winding, like for a 5Y3 or 5R4, swapping the rectifier out for a 5U4 (3A) or the 5V3 (3.8A) it could, over time, damage the power transformer. When we examine the ethereal level amplifiers from long ago, I am talking about 1940 or so, the output was SE and the harmonic profile was that of a triangular wave with all harmonics in a harmonically descending amplitude, which, as far as analytical measurements go, is horrible... but that is the sound that we like music created with. I am also convinced that we also like the small artifacts created in the vacuum tube amplifier itself. Absolute, pure and clean, 0.001% THD amplifiers tend to sound sterile to most of us.

@@ElPasoTubeAmps You're right, it was a 5R4 now that you mention it. Wish I could get more of those at the price they were going for in the 90s. Another thing I realized while recalling my days of similar experiments, is that aside from the filament current, you also have to watch out for the value of that first capacitor hanging off the rectifier. You can put as much capacitance as you want after a resistor or choke, but that very first cap must not exceed the max capacitance value for that tube or you can kill the tube with all the inrush current. For guitar amp experimenters, this could be a real issue if you had a 5AR4 originally and the max permitted 60uf capacitance, then you switched to a tube like 5U4 which only allows 40uf.

You are referring to the so called "potato masher" version of the 5R4; I believe they were made by Chatham Electronics as well as possibly Cetron.

I have 2 of those big yellow 5R4WGY's....The box's say April-1955 . I read the WGY suffix was for military spec, somewhere. I tried them in my old Ampeg '65 Portafex B18-X head. Spec'd for a 5AR4. Much more sag with the 5R4. I also have a ss rectifier , but the 5AR4 sounds the best. I have both the 5R4WGY's and boxes they came in if anyone wants pix of them. Not for sale, though... :)

@@goodun2974 You are correct, mfg. by Chatham Electronics in Newark, N.J.

@Kevin Counihan, interesting idea, thanks!

For lowest distortion and highest output across the entire power range of the amplifier, it is like you say, solid state rectifiers that cost nothing and last forever, would be the choice. On the other hand, me, not being a musician, I can't say about the sound produced by the sag in voltage across a high-vacuum rectifier. I listen to this particular amplifier on a daily basis and use a 5U4 in it because it is a center channel and probably never delivers over 5-10 watts. It provides a good 50 watts with the 5U4 and about 80 watts with SS rectifiers. I suppose if it were to be used in a venue where the highest amount of power was required, I would plug in the SS rectifier. Not sure all that I said in the video but a 5U4 category rectifier is really only good for an amplifier that provides about 30 watts output. Off the subject of this video - that is why I sometimes use mercury vapor rectifiers. They are the 100 year-ago solid-state-performance rectifiers. About 15 volts drop no matter the current draw and even the small ones (816 MV rectifier) can power a 400 watt transmitter.

@@ElPasoTubeAmps Thanks for a bunch of interesting info. I use two 5u4gb in my DAC output. One for each mono channel. The 6n2p triode draws about 50mA so not exactly stressing these guys. I built it using them because I like the look of tubes, and guys like Lukas Fikus, aka the Lampizator swears by them. I have been using the same NOS Marconis for 10 years. Replaced them today with the cheap Electro Harmonix . The new tubes certainly brought the dynamics and punch to life.

@@user-od9iz9cv1w I love tubes too and the lower power tubes like the 12A_7 family of tubes can last virtually forever. On the other hand power tubes may not last so long. 50 mA is sure not stressing your 5U4's. Notice in high-end amplifiers like the McIntosh MC60, they use two 5U4's for 60W output. In the MC30 there is only one 5U4. I have found out the hard way... what I mean by that is a lot of detailed trouble shooting with no results in solving a problem... that in power circuits where the level of DC current of the circuit is in the upper portion of the maximum current (for example) a 5U4 can supply, can cause issues hard to diagnose. I am that type that wants to find the "real problem" and not just swap out tubes but this has caused me unnecessary work at least once when the 5U4 is weak. Best example of this happened in a Collins 32S-1 transmitter where all the voltages were within specs on the bench but not when it was transmitting and the load was the highest. I tinkered on and off with this problem for probably a year... The trouble shooting was done on the work bench but for some time, I failed to test the transmitter into a dummy load under full power output. I only mention this as sometimes a little bit of careful tube-swapping (with a known good power tube or rectifier) can reveal a problem much easier than chasing and smashing electrons as we may like to do. 🙂 Stay safe.

Eimac has an article on "cathode stripping" and they say it only occurs at extremely high voltages - 50KV or more? so, I don't think it applies to our world. As for the 5R4, I do see them used occasionally in homebrew amps but they are for higher voltages and lower currents than the 5U4 so odds are, being used in a small audio amp, the 5U4 would be a better choice than the 5R4. The Collins 30L-1 amplifier used a 5R4 to deliver 800 volts to the plates of the 6146's and a 5U4 for the rest of the equipment like the KWM-2A or 32S transmitters.

@@ElPasoTubeAmps Agree and thanks for additional info. I can check the 5R4 tube characteristics and get some rough values but there's nothing like the real thing. The reason why I brought up 5R4 was back in time, people pushed 5U4 with their home brew xmiters to the point where they started arcing. 5R4 came to the rescue. So glad to find your channel. Thanks and keep them videos a coming!

A $1 thermistor can solve the slow warmup issue.

I think the 5AR4 has a reasonable "slow warm-up" characteristic but the 5U4 is hot and conducting long before the 6SN7's and KT88's are conducting so voltage does soar above the steady state value with the 5U4. Even the 5AR4 is a bit quicker to conduct and has been seen as a potential issue in the Dynaco MarkIII amplifiers where the turn-on DC voltage exceeds the 525 volt rating of the electrolytic capacitors in that amplifier before the PS is properly loaded by the output tubes (6550's). As for listening, this amp sounds "big" and powerful to me. The smaller amps seem to have more detail but the bigger amps just sound powerful from the get-go. Just my impression of listening to amps.

I really agree. Have not been impressed with solid state rectifier conversions. Nice to have the tube! Really interesting to see the performance variation -- the 5V4 is a bit of a sleeper...

@@HiFiTown I agree that the 5V4 is a nice rectifier. The 5AR4 has a current rating a bit higher (250 mA) to the 5V4 (175 mA) but if not over loaded, the 5V4 would do just as good a job and probably for less money. A current Ebay listing for the 5V4 - www.ebay.com/itm/Vintage-Lot-Of-4-NOS-RCA-5V4GA-Rectifier-Tube-USA-Vacuum-Tubes/202815936238?hash=item2f38c596ee:g:JU8AAOSwFWVdwYnk

I had a 5AR4 a few years ago that was too simple a problem but took me a long time to figure out. Only half of the rectifier was working. I let the easy fix go over my head and afterwards felt dumb for it taking so long. Also, in my Collins radio equipment, the PS runs a 5R4 for HV and 5U4 for LV. I have had issues there where I swap the tubes and basically nothing changes but after time and maybe putting in a SS replacement, it works properly and I find out that several of my old tubes are weak. Yes, as simple as it may seem, vacuum tube rectifiers can bite us really hard sometimes. I will mention another PS problem that turned out to be a real surprise. If a PS has a choke in it and the capacitor on the output of the choke opens up so there effectively is no capacitance there - all will appear OK with the amplifier just sitting turned on (DC voltages are fine) but once you start driving it, it will get terribly distorted. The reason for this is that capacitor on the output of the power supply choke takes the signal from the power supply side of the output transformer to ground and if that capacitor is not there (open) the driving signal you are putting into the amplifier can not get to ground thru the power supply choke. I hope it makes sense. The DC component (the steady state voltages) in an amplifier work separately from the AC signal (not the power line AC) component of the signal path. Capacitors with one side to ground not only take the ripple from the PS to ground to clean it up but it also puts some of the components (especially the PS side of the output transformer or CT of a push-pull amplifier) to ground to take the driving signal to ground. PS problems often seem simple but it is easy to overlook them sometimes. Hope this helps.

@Stephen Morton It's not 6 V lower it's 60. Could also be the regulation of the transformer from no load to full load. Need to monitor the ac into the rectifier as well to see where the drop is. Not convinced it's the rectifier alone, my main suspicion is the filtering though.

@Stephen Morton sorry, didn't see the decimal dot in front of the 6, even more reason for my question.

I will see if I can measure ripple, idle and with full load. I can say that the S/N of this amplifier is about -89 dB so it is extremely quite. I did measure that parameter.

@@ElPasoTubeAmps Thanks David. As you're probably aware the ripple is less of a problem in the push pull OP stages , I presume you have additional filtering in the low level stages which is likely the reason for the good low noise performance. Reason I mentioned the filtering is that I have seen the effect of it on some of the old Fender guitar amps few years ago. I just did a similar test on an old Mesa Mk IV head, this thing has a massive mains transformer and a pair of 220 uF 300 V electros in series totalling at 110uF running off a centre tapped 690 - 700 Vac via 2 seriesed pairs of half bridge 1N4007 diodes, pretty lame choice of diodes in my opinion! The B+ sags by about 34 Vdc from no load to point of clipping on the output, the ripple also increases but no where near as much as Ii remember in the Fender amps. Measurement of the AC to the rectifier diodes stays virtually unaffected! I would have liked to swap out the the 4007 for a set of 1N5408s but as this is a customer repair not an option fr number of other reasons as well. I can't recall if you mentioned which SS diodes are used in your amp?

Drop across MV rectifiers is constant as I understand it. Looking at the ratings, the 816 is good for 1600V at 250 mA (400 watts) and the 866A is good to 3200V at 500mA or 1600 watts. Pretty impressive for 70+ year old technology and tubes that still work... See page 20 of this document and notice the prices of the 816 and 866. www.nj7p.org/Manuals/PDFs/Tubes/RCA-gfat-1942.pdf

@@ElPasoTubeAmps 1600V at 250mA is very good, I've mainly used 866's my whole career. There are other larger MV rectifiers that are used in the larger AM transmitters that put out some serious B+. I've repaired a few RCA 25 and 50KW xmttrs that used 872A's. That is a great old data booklet that RCA put out. Contains a lot of valuable info on rectifiers, and transmitting triodes. There's a lot of tubes I see there that can be used for audio. Thank you for turning me on to this booklet. It is now in my file folder of tube data :)

@@djfrank59 I was actually considering 872A's but the price on Ebay is just too much and they aren't necessary. I am really pleased to discover the 816's. That could be a fun amplifier to build and operate.

I agree with you. Apparently, Telefunken buys JJ's very best and puts their name on them as all four of mine are exactly the same and have remained so. A little pricey but can be worth it in a nice high end amplifier. Thanks for your comments.

@@ElPasoTubeAmps for sure! The JJ 88s are remarkably consistent and good quality IME, about as consistent as the much more expensive Gold Lions out of Russia.

I see the 159v is rated at 1.5H and 500 mA - that should be good. If your 420 volt winding is CT and you use it in FWCT, you should get around 275 VDC. In a bridge you would get twice that much.

hi no, i got 420- 0-420 volts on the power transformer and i put a 10 h choke before the 2 chokes 159v to separate the 2 channels of the amp.@@ElPasoTubeAmps

I see there have been several discussions about drop across the choke and loading of the transformer, etc. As for the choke we can do some calculations. DC resistance of the choke is 120 ohms so at idle current of 150 mA we have 18 volts drop and at full current of about 270 mA we have 32.4 volts and so on. Factoring in the DC resistance of the secondary of the transformer I think it will all add up nicely. I thought about a meter across the rectifiers but then decided that would be just too much data. No doubt the SS rectifier is constant at about 0.7V and each of the vacuum rectifiers has its own characteristics which all together, makes it turn out what it does. Possibly considering the voltage drop across the SS rectifiers constant, we could estimate the R value of the secondary of the transformer as the difference in voltage drop from idle to full current (i.e. 270 mA - 150 mA) = 120 mA and a voltage drop of approx 60 volts and come up with a R value for the secondary of about 500 ohms minus the choke R which would come out to around 380 ohms (?) Does that sound reasonable for a transformer rated at 250 mA? Seems a little high in R to me. Anyone's thoughts?

It would be superior and provide a little more power output from the amplifier and those small gains were what the guys were looking for back in those days. Just a little gain...but as life seems to be, the gain in less voltage drop from a vacuum rectifier is offset by the fact that the MV rectifier, 83, has to warm up before it will work. P.S. they use the 83 in the TV-7 tube testers for the constant and low voltage drop across the rectifier. MV rectifier of their day were the solid-state rectifiers of nowadays. Speaking of warming-up... it seems that today it takes more time for our "computers" to "boot" up than it did for the tubes to warm-up. Think of the aggregate time in a life-time spent waiting for a computer to boot up compared to tubes warming up.

That is a good question. I would say for one thing, people just like tubes. Secondly, playing guitar amplifiers with a vacuum tube rectifier in it has "sag" which means under maximum output power conditions, when the amplifier is drawing maximum current thru the rectifier, the voltage drop across the vacuum tube rectifier increases, ending up decreasing the available voltage to the output tubes and they distort or "sag" in a way that the musician tends to like in creating music. Solid state diodes have a fixed voltage drop regardless of the current thru them of about 0.7 volts each. Depending on the current an amplifier is drawing thru a vacuum tube rectifier, the classic one being a 5U4, the voltage drop across the 5U4 could easily be forty (40) volts. This would from an engineering point of view, be seen as a loss and something we want to prevent but in the hands of a musician, it (may be) musical creativity. This is a reasonable explanation of why many musicians like vacuum tube rectifiers. As for actual all vacuum tube vs SS amplifiers (tubes from the input to the output or SS from the input to the output), the sound of a vacuum tube amplifier driven into distortion has a much more "rounding" of the distortion products, which is not so harsh to our ears, as a SS amplifier will be when maximum (and more) output power is obtained. SS amplifiers distort with more jagged edges on the output waveform and these sharp corners on the distortion products are not pleasant to our ears. Hope this helps.

I have studied and built and modeled these resonate power supply circuits with LTSpice and there is more to them than meets-the-eye, so to speak. Think about this - if the LC circuit was perfectly resonate at USA frequencies of 120 Hz output from a full-wave rectifier - what would the parallel circuit look like at resonance? It would become an extremely high Z circuit and all the voltage would be dropped across it - that is what would happen, so we really don't want "perfect" resonance of this input power supply circuit. If you study real-world power supplies like in Collins and Gates transmitters and do the math for the values they use, you will see that they are calculated to be resonate a little bit above 120 HZ. And, with that said, as current increases, the L will start to drop, maybe only slightly, but it will drop and resonance will rise - but - initial resonance is placed slightly above 120 HZ so that "perfect" resonance will never go thru the 120 HZ point with changing values of L under load. It would be a disaster if, under modulation load for example, the current thru the choke was making the LC circuit be resonate at 120 Hz each time it passed thru that point from fluctuating inductor value under load. Take a look at the Collins commercial transmitter in the link below. The schematics are at the very end. The original values are 10H and 0.15uF which gives Fr = 130 HZ. Someone has penciled in 20H and 0.08uF which gives Fr = 126 Hz. You will see the same thing if you study Gates AM transmitters. www.steampoweredradio.com/pdf/collins/manuals/collins%2020v%20am%20transmitter.pdf

@@ElPasoTubeAmps Thanks for the link. It will improve voltage regulation by reducing voltage at low load. Yes it will need to be tuned. The chokes they use are called swinging chokes.. It becomes more like capacitor input under heavier load.

@@mohinderkaur6671 Chokes used with shunt capacitors for resonance are not swinging chokes. As I am sure you know swinging chokes are high inductance, something like 25H at low current and 5H at their maximum rated current. I have a couple of these and I have tested them under varying loads and they perform as advertised. I have never used the swinging chokes in power supplies I have built. The value of the swinging choke comes from the idea that, as a rule-of-thumb; L (critical) = Vdc/Ima. Here is an article that discusses choke input power supplies. ken-gilbert.com/choke-input-power-supplies-part-1-henry-pasternack/comment-page-1 Their formula says that for a 3000 volt supply and a 10H choke we will need a drain of 300mA of current to maintain a reasonably stable output voltage but at 25H we need only 128 mA of constant current for reasonable stability in our output voltage. As they mention in the article, solid state devices have evolved to the point that they can take the high current surge straight into a capacitor only filter. I still use chokes as I have them and I like them and they do provide better regulation. I have a 1947 Collins transmitter (30K-1) that uses a resonate capacitor/choke input to a second choke as a Pi and the output voltage running at 3200 is amazingly stable under modulation. It seems the resonate choke input followed by a Pi might be the best. As I mentioned, I modeled these power supply components several times in LTSpice and I never got a satisfactory answer/understanding in simulation. Design and simulation is fun but generally I check known good designs like Collins Radio, and others, for what has been proven to work and start there instead of trying to reinvent the wheel. 🙂 Thanks for the discussion.

@@ElPasoTubeAmps I used these to reduce 380-0-380 volts from transformer rectified by 5u4g down to more like 330v dc for EL84 push pull stereo amplifier. Thanks for the link.

I don't take in equipment for repair anymore but you have a nice instrument in the L4B. A simple but beautiful amplifier if you feel comfortable working with 2500 volts but if not, I would not attempt it. Could you briefly describe the issue you have with it? My direct RU-vid email is : theaudioshop@sbcglobal.net

It is the closest one for sure. As mentioned in a comment above, the spacing of the plate and cathode in the 5AR4 is very close and does allow for the least amount of voltage drop. I like the 5AR4. If a person just loves the vacuum rectifier and has to have it in an amplifier, I can see wiring it in and then underneath on the tube socket, shunting the plate to cathode with SS diodes and have the best of both worlds - beauty and performance...

I thought about the 83 as I do have some but the current rating is only 125 mA for that little tube and I am concerned it might just be damaged. I think it would be a fun rectifier in a smaller amp with maybe a pair of 6V6 at 15 watts or so.

It was just a thought because it has very low Drop in voltage and I've always gotten very stable power output out of it. I currently have one in a 2 6 L 6 amp making about 40 Watts but you are doing something considerably bigger. Also the reason I mentioned it is you tried the 5y3.

@@kevin2960 I like the idea and may use the 83 in a future video. It is used in the TV-7 line of tube tester for its constant voltage drop so that is a very nice advantage of the MV rectifiers. Thanks for the information on how you are using it in your amplifier.

@@ElPasoTubeAmps , it would be simple to make an adapter so that the type 83 tube could drop into a 5U4-wired octal socket. Regardless of the official tube- manual rating for the type 83 tube, it may interest you to know that a short-lived version of the mid to late 50s tweed Fender Bassman actually used a type 83 rectifier, and although I havent seen or worked on that version ( I wonder how many were actually manufactured?), based on my experience working on Fender amplifiers I can pretty much guarantee that the rectifier tube would have been forced to deliver more current then it was rated for! By the way, I did once repair a Bassman that had *most* of the circuit design of the type 83 version (as shown in a factory schematic), with some circuit changes to reflect the next "official" circuit and schematic to come, but it had a factory-original octal socket for the rectifier tube ( 5U4 as I recall). Fender amp designs were always in a state of flux anyway, and it is fairly common to find specific amps that deviate from the factory schematic for that model-year.

@@goodun2974 I completely agree with you on running tubes beyond what the vacuum tube manuals specify. Their data seems to be very conservative and probably a CCS rating especially the receiving tube class of vacuum tubes As for the transmitting class of tubes and the big MV rectifiers, I run a pair of 866A's in the PS of my RF amplifier in a voltage doubler. That is definitely a no-no as the HV MV rectifiers are never seen operating into a capacitor but rather always into a choke input filter - probably a two stage with a swinging choke into a Pi filter. But it works for me and I turn the PS on each time I transmit and off each time I quit transmitting so it is on and off all the time and the output is 3.5 KV. There is also a lot of fear-mongering about "hash" and all kinds of noise from MV rectifiers. I have never experienced it and plan on using some 872A rectifiers in my next RF project. I am not saying that some common knowledge and common sense is not valid but I am going to push the parameters a bit.

The early rectifiers including MV's all had the same pin-out...right through the 60's/70's. Octal, or 4 pin. Theoretically, , the higher power rectifiers used the 4 pin which was a little safer. The larger pin spacing prevented arc-over from pin to pin to prevent PT damage. and most of the time, even on the octal-based rectifiers, only 4 pins are used. Also, the larger pins offered better current carrying capability, with less resistance and a tighter connection. Some of the smaller high-vacuum rectifiers had a filament center tap to power the pilot light on many AA5 radios. usually the 35 volt tubes like 35Z5 etc. Rectifier tubes always worked the hardest in radios and TV's If you look into the design of earlier tube TV's, many of the designs used to have the rectifier tube 5U4GB mounted in a vertical chassis and the tube was in a horizontal position which shortened the life of the tube due to filament sag which eventually caused it to collapse and short to the plate. That design didn't fly well in those sets. many PT's were taken out because of it. Very few sets back then had a primary fuse to protect the PT.

@@djfrank59 Nice history there, Frank. thank you. I really enjoy learning from your guys experience and I never actually thought about the layout of the pin configuration on the rectifiers. Makes sense to separate them as they are... I worked a few years in TV repair shops now, more than 50 years ago and I determined and always told people that the big TV sets of those days basically just burned themselves up from heat. No fan in them and might stay on all day. From what I remember, things got even worse when TV's started being build on PC board. What a disaster for the cheap boards and, again, the heat. Besides crystallized solder joints on the sockets from the heat, I remember one in particular where the horizontal output tube was the type without the plate cap and I worked on it so long with a red-plate condition and it ended up being the tube socket pin for the grid, if I can accurately remember that long a go. The point here is, we have talked a number of times over my videos how open tube socket pins (especially the plate connection) and how an open plate condition will instantly ruin the screen grid in the output tubes but in the case of the open grid pin, there was no bias to control plate current so a similar type condition but at least it was the plate and not the screen being vaporized. Funny how I still remember that.

@@ElPasoTubeAmps I did TV/Radio repair also for many years as well, and Horizontal Output/Damper tubes always cooked as it was, but just like a guitar/hi-fi amplifier, if you lose the bias, the tube cooks. The common HOT's were the 6DQ6,, 6BG6, 6BQ6, and for color sets, it was the 6DQ5, 6JE6, and 6JS6. But many of the later Compactron tubes used for Horizontal sweep didn't have an external plate cap. They also contained the Damper in one envelope, and I've seen many arc-over situations. But the worse design mistake in TV chassis design was the Vertical chassis. The Rectifier tube was always at the bottom and mounted horizontally which made the excessive heat rise to the other tubes which shortened the lives of the other tubes as well as resistors and caps. Sweep tubes in a set always work the hardest and had the shortest life, next to rectifiers. I still repair and restore a lot of tube TV's for collectors and the PCB designs are the worst! The Philco Predicta series sets were the most unreliable as was early color sets. I always find burnt tube sockets and boards in those sets. Design engineers pushed a pencil and didn't have to repair these sets, so they catered to the bean counters at the companies..they didn't care about the repair technicians or the buying public.

@Kevin Counihan I've seen not only glass melt, but holes in the glass as well. even the glass splitting in 2 pieces. I deal with it more now than ever. today's guitar amps are run beyond their limits. When you see an EL-84 biased a 40 mA, you know where that's heading LOL I still do repairs.

Indeed... if each component is purchased separately new and some off Ebay, I think the amp would cost $600-$700 in parts to build.

What a sentence.